Physiology of Stress Tolerance

in Rice

Proceedings of the International Conference on Stress Physiology of Rice

28 Feb - 5 March 1994, Lucknow, U.P., India

1996
NDUAT, IRRI

The International Rice Research Institute (IRRI) was established in

1960 by the Ford and Rockefeller Foundations with the help and
approval of the Government of the Philippines. Today IRRI is one
of the 16 nonprofit international research centres supported by the
Consultative Group on International Agricultural Research (CGIAR).
The CGIAR is sponsored by the Food and Agriculture Organization
of the United Nations (FAO), the International Bank of
Reconstruction and Development (World Bank), and the United
Nations Development Programme (UNDP). Its membership
comprises donor countries, international and regional organizations,
and private foundations.
IRRI receives support, through the CGIAR, from a number of donors
including FAO, UNDP, World Bank, European Union, Asian
Development Bank, Rockefeller Foundation, Ford Foundation,
Common Fund and Commodities, Sasakawa Foundation, and the
international aid agencies of the following governments : Australia,
Belgium, Canada, Colombia, Cte d Ivoire, Peoples Republic of
China, Denmark, Egypt, Finland, France, Germany, India, Indonesia,
Islamic Republic of Iran, Italy, Japan, Republic of Korea, The
Netherlands, Norway, Philippines, the Russian Federation, Spain,
Sweden, Switzerland, United Kingdom and United States.
The designations employed in the presentation of the material in this
publication do not imply the expression of any opinion whatsoever
on the part of IRRI concerning the legal status of any country,
territory, city, or area, or of its authorities, or the delimitation of its
frontiers or boundaries.
International Rice Research Institute 1996
Los Banos, Philippines
Mailing address : P.O. Box 933, Manila 1099, Philippines
Phone :
(63-2) 818-1926,812-7686
Fax
: (63-2) 891-1292
Email
: Postmaster @ IRRI.CGNET.COM
Telex
: (ITT) 40890 Rice PM; (CWI) 14519 IRILB PS;
(RCA) 22456 IRI PH; (CWI) 14861 IRI PS

Suggested Citation :
Singh, V.P. et al. (eds) 1996. Physiology of Stress Tolerance in Rice : Proceedings of the
International Conference on Stress Physiology of Rice, 28 Feb - 5 March 1994, Lucknow,
U.P., India, 239 p.

ISBN NO. 81-86789-00-6

CONTENTS

Sl.
No.

Particulars

Author (s)

Page
Nos.

FOREWORD
ACKNOWLEDGEMENT
INTRODUCTION
1-10

1.

Improving cereals for the

variable rainfed system : From
understanding to manipulation.

K.S. Fischer

2.

Is there anaerobic metabolism in

submerged rice plants? A view
point.

H. Greenway & TL. Setter

11-30

3.

Tolerance to submergence and

anaerobiosis : Rice seedlings
have something still to learn.

Michael B. Jackson, Deborah

M.E. Peare & Jacky E.
Summers

31-43

4.

Environmental and plant

measurement requirements for
the assessment of drought, flood
and salinity tolerance in rice.

P.C. Ram, B.B. Singh, A.K.

Singh, V.K. Singh, O.N.
Singh, T.L. Setter, R.K. Singh
& V.P. Singh

44-69

5.

Nutrient uptake by roots of

cereals in oxygen deficient
environment.

M.C. Drew

70-79

6.

Response of flooding in
gogorancah rice and moisture
stress effect at reproductive
stage in walik jerami rice.

Didiek Setio Budi & Bambang

Suprihatno

80-90

S1.
No.

Particulars

Author (s)

Page
Nos.

7.

Anaerobic tolerant rice seedling

growth and its potential in the
tropical rainfed lowland
ecosystems.

M. Yamauchi, A.M. Aguilar,

J.L. Padilla, D.V Aragones &
U. Singh

8.

Carbohydrate status of rainfed

lowland rices in relation to
submergence, drought and
shade tolerance.

G.S. Chaturvedi, P.C. Ram,

A.K. Singh, P Ram, K.T.
Ingram, B.B. Singh, R.K.
Singh 6 V.P. Singh

103-122

9.

Stress physiology in relation to

breeding for drought resistance
A case study of rice.

S. Fukai & M. Cooper

123-149

10.

Varietal characteristics and soil

management to reduce drought
stress.

H.U. Ahmed, M.L. Ali, S.K.

Zaman, M.A. Kabir & N.M.
Miah

150-167

11.

Modelling the soil-plant-water

balance of rainfed lowland rice.

M.C.S. Wopereis, M.J. Kropff

& T. P. Tuong

168-185

12.

Drought response of dry seeded

rice and its modelling.

A.K. Singh, T.P. Tuong, K.T.

Ingram, R. Pernito & J.
Siopongco

186-197

13.

Rice root systems development

under rainfed conditions.
Social consequences of stressed
environments.

G. Pantuwan, K.T. Ingram 6

P.K. Sharma
Thelma R. Paris, Abha Singh
& Mahabub Hossain

198-206

Differential growth in rice

production in Eastern India :
Agroecological and socioeconomic constraints.

Mahbub Hossain & Alice

Laborte

14.
15.

91-102

207-220
221-239

FOREWORD

The needs of the rainfed lowland rice ecosystem in Southeast Asia are
enormous. The hundreds of millions of the worlds poorest people living in this
region depend upon the almost 40 million hectares of rainfed lowland rice. The
extent of the area combined with a difficult, heterogeneous, and variable
environment demands that a new research approach be adopted. The Rainfed
Lowland Rice Research Consortium was established in 1991 to provide a framework
within which national and international institutes could bring their resources to bear
on this important, though difficult environment. National programs from
Bangladesh, India, Indonesia, the Philippines, and Thailand joined with IRRI, and
supported by the Asian Development Bank, to develop a coordinated research
strategy and program directed towards improving the sustainable productivity of
this environment.
This conference on stress physiology represents several important steps in the
growth of the Consortium. First, it is directed towards assessing the status of strategic
understanding of two of the most critical constraints in the ecosystem: drought and
submergence. Research investment has been increasing in these areas, yet there are
few fora in which scientists can exchange views and findings while focussing on
this ecosystem. A systematic compilation of the present status is essential to careful
planning of future steps.
Second, scientists from Consortium member countries were able to interact
with scientists from advanced research laboratories from Europe, North America
and Australia and begin to plan joint "North-South" research and collaboration.
Scientists from other countries with important rainfed lowland production areas
were able to attend, even though their institutional strengths do not yet permit full
Consortium research responsibilities. It is expected that the contacts made at this
and other similar conferences will stimulate "South-South" collaboration.
Third, lively and vigorous discussions were carried on throughout the event;
both in plenary sessions and in the corridors. The active participation of members
from all the countries and institutions represented indicated that the spirit of
scientific research has permeated the culture of the Consortium. This collegial spirit
is essential if true partnership is to guide research in this complex ecosystem.

Finally, the event was hosted by one of the member institutes, Narendra Dev
University of Agriculture and Technology, with support from the Indian Council of
Agricultural Research, and the Uttar Pradesh Council of Agricultural Research.
This marks an important step in assuming the full range of responsibilities
that come with Consortium membership. It is fitting that this event took place in
the midst of such an important rainfed lowland rice growing area. That the opening
was graced with the attendance of the Governor of the State of Uttar Pradesh is an
indication of the importance that local governments place on improving their
agricultural productivity. Indeed, if Uttar Pradesh were a separate country, it would
be the 7th most populous nation in the world.
I congratulate the participants and the organizers for realizing this event.
These proceedings will serve as a valuable reference for scientists working in this
dynamic field. It is especially significant that, through the Consortium, this product
will be readily available to scientists working in developing countries where access
to the latest physiological literature is severely limited. The host country and the
host institution have made this event a very memorable one. The scientific
competence and a specially warm hospitality has impressed each and every
participant. I am pleased that IRRI working with Consortium members has had the
opportunity to contribute to this important event. I sincerely hope that there will be
many more in this tradition in the future, and that IRRIs role will continue to evolve
towards that of an equal participant among colleagues sharing a common goal.
Klaus J. Lampe
Director General, IRRI
12 September, 1994

vi

ACKNOWLEDGEMENT

Several scientists and organisations made valuable contributions to this book,

which without their help would not have been possible. Special acknowledgements
go to the contributing scientists who embellished it with their precious time,
experience and ability. The organisations which supported the scientists while they
worked on this project merit myriad commendations and thanks. Valuable
suggestions given by various reviewers are thankfully acknowkedged. In this
context, the name of Dr. G.C. Srivastava, Principal Scientist, Division of Plant
Physiology, Indian Agricultural Research Institute, New Delhi deserves special
mention.
The editors are specially indebted to Dr. P.C. Ram, Dr. G.S. Chaturvedi, Dr.
R.P.S. Chauhan and Dr. A.K. Singh of N.D. University of Agriculture and Technology
(NDUAT) for their assistance in editing the papers. We are thankful to Dr. Maharaj
Singh, Mr. H.P. Singh and Mr. S.P. Singh of the Department of Crop Physiology,
NDUAT, for proof reading the manuscript.
Several people in the Computer cell of the NDUAT, have contributed much
through their secretarial, technical and artistic talents. The editors fully appreciate
their contributions which resulted in improving the book considerably.
Editors

vii

INTRODUCTION

The rainfed lowland rice ecosystem covers half the rice growing area in the
humid and subhumid tropics of Southeast Asia. Including the contiguous
flood-prone areas, almost 50 million hectares are sown to rainfed lowland rice. The
system is characterized by environmental heterogeneity and variability, with
frequent droughts and floods damaging the rice crops. Modern rice technology has
barely contributed to rice production and, consequently, yields remain low, in many
areas less than 2 tons per hectare. None the less, this ecosystem provides the staple
food for hundreds of millions of the worlds poorest urban and rural dwellers. But,
as populations and economies continue to grow, increasing pressure is placed upon
the rainfed lowlands to help meet the estimated 70% increase in rice demand
projected over the next 30 years. This pressure is even more acute as investment in
irrigation declines, and water and land are diverted from irrigated rice agriculture
to urban and industrial uses.
Research is required to increase the productivity of this vast ecosystem, and
this increase must be sustainable. That is, the sometimes conflicting demands of
social and economic development must be reconciled with husbandry of our social
and natural resources. Developing technologies that permit such productivity gains
while preserving or enhancing the resource base will demand a fundamental
understanding of the processes that detrermine plant adaptation to the environment,
the dynamics of water and nutrient flow through the system, and the interaction of
management practices designed to maximize efficiencies with the biotic, social, and
economic environment.
The distribution and enormity of the rainfed lowland ecosystem combined
with the complexity of the environment and the difficulties of the production
constraint required that a new research approach be adopted. In 1991 the Rainfed
Lowland Rice Research Consortium was formed. It is a formal linkage among
institutions in five countries (Bangladesh, India, Indonesia, Philippines and
Thailand) and the International Rice Research Institute. The member institutions
agreed upon a research strategy specific to achieving sustainable productivity
increases in the ecosystem, and apportioned responsibilities according to
comparative advantages and available infrastructure and personnel. The Rainfed
Lowland Rice Research Consortium certainly represents a paradigm shift from a
target of uniformly favourable environments addressed from a centralized research
viii

effort, to a decentralized effort involving many institutions at all stages of research

and targeting more difficult environments. In effect it provides a mechanism to move
the needed strategic research from experimental centers located in irrigated
environments to centres and farmers' fields situated within relevant environments.
Although most of the resources dedicated to this research efforts are from IRRI and
the participating national systems, supplemental research support has been provided
by the Asian Development Bank in the form of a Technical Assistance grant.
The overall research strategy adopted for the rainfed lowlands is based on the
development of adapted germplasm that can tolerate the prevalent stresses. With
such tolerance it is then possible to develop responsive cultivars that, under higher
nutrient status, can produce higher yields. This in turn opens a range of alternatives
for farmers to diversify and intensify their production systems. While developing
stress tolerant cultivars is not sufficient in itself to reach our goal of sustainable
productivity increases in the rainfed lowlands, it is an essential component in the
strategy.
Recognizing the importance of understanding the processes that determine
adaptation and stress tolerance, the Consortium sponsored this conference to
establish the "state of the art" in the physiology of stress tolerance in rice as it relates
to drought and submergence. 75 scientists from 11 countries were invited to Eastern
India, an area where both stresses are found and where a major research effort is
underway within the Consortium. Participants were asked to present their best
assessment of what is now known regarding drought and submergence tolerance,
and plot a course for future research. We also asked that they assemble for informal
brainstorming sessions to help the Consortium plot its future directions. By bringing
scientists from advanced research centres in Europe, North America and Australia,
it was hoped that collaborative projects meeting the needs of the Consortium and
of the research laboratories could be developed. Finally, we included an assessment
of the social environment in which farmers work and live. It is essential that
process-oriented research be anchored within the social context for its output to be
developed into relevant applications.
It is impossible to summarize here all the discussions and views that emerged.
Since many points are very thoroughly covered in the papers, we will only touch on
particularly strategic issues that were raised. Perhaps the most persistent theme that
emerged from the discussions was the need to better understand the target environments
in which we are working. Without suitable descriptions it is impossible to interpret
experimental results and generalize or extrapolate from even the most strictly controlled
studies. At the very least standard environmental descriptive data must be collected for
every experiment to allow comparison of results among sites. The very nature of the
stresses encountered and the environmental factors that contribute to severity and
modify plant response are still poorly understood (e.g., interaction of water and nutrients
in water- deficit conditions, and the interaction of turbidity, cloud cover and water
temperature under submerged conditions). More complete descriptions and
quantification of environmental variability are needed in oder to better understand
how very different physiological processes may be confounded into similar phenotypic
ix

responses. The probabilities of stresses of given severity should be determined in

order to develop extrapolation domains of tolerances at given levels. "Probe
genotypes" should be selected that will aid in quantifying differences due to
environment and genotype by environment interaction using replicated trials.
Research in a consortium mode, with a number of teams spread throughout
a vast geographical region and working within diverse institutional environments
can allow a proliferation of research activities that lose focus. Thus, experimentation
should be restricted to a relatively small number of key experiments that are tightly
linked to the principal themes being addressed by the Consortium. The development
of multi-site experiments on key problems will be a major contribution both to
environmental characterization and to understanding environment by physiology
interactions. Concentrating on a set of such experiments will help us retain focus and
will contribute to the development of a team approach to these difficult problems.
Standardization of methodologies will be a direct result of developing
cross-site research . Current research output suffers greatly from the difficulty of
directly comparing outputs from different sites. In order to establish a common
working mode across the environments and across continents an effort will have to
be made in training scientists in standard methodologies. Agreement should be
reached on what particular measurements must be taken, and how. Clearly
minimum data sets for particular kinds of experiments must be developed. These
will include key weather/climate and soil/hydrology parameters. These agreed
standards will contribute to our efficiency and, of course, will be more
cost-effective. The development of this team approach is a direct consequence of the
Consortium approach, and is an indication that the attempt to join forces can promote
the integration of sound science at higher levels.
The rainfed lowlands still lack a clear articulation of "ideotypes" for target
environments. This may simply reflect the previously stated need to clearly articulate
our target environments. The ideotypes should incorporate adaptation to various
timings and severity of stress, as well as major crop establishment alternatives. It is
very likely that large rainfed lowland areas will convert to dry direct seeding, and
that this will have a significant impact on tolerance to and management of early
season drought. What plant types are most appropriate for this system, and where?
Finally, there was a very strong expression of need for greater links with social
sciences and a need to incorporate a social science perspective in strategic stress
tolerance research. It was felt that early in the research process our efficiency could
be increased by understanding how social constraints contribute to technology
adoption, or lack of adoption. Risk reduction is certainly a driving force for
improving stress tolerance, but how is it perceived, measured, and managed by
farmers today?. The central role of women in the management of rainfed lowland
rice system is often ignored, with potentially serious consequences. The very low
representation of social scientists in consortium teams and the relative lack of social
issues within the research agenda were seen to be weaknesses of the present structure
and approach.

There is little doubt that research into the physiology of stress tolerance can
make a major contribution to improving sustainable productivity of rainfed lowland
environments. However, with the finite resources available to us it is essential to
sharply focus our efforts and take full advantage of the human resources available
to us. Fragmentation of effort and a diffuse approach along more academic lines will
simply delay our progress and frustrate our efforts. This conference has been an
important step in developing the team approach and, if followed up by appropriate
planning and training, should lead to rapid progress over the coming years.
Editors

xi

IMPROVING CEREALS FOR THE VARIABLE RAINFED

SYSTEM : FROM UNDERSTANDING TO
MANIPULATION
K S. Fischer1

SUMMARY
Inspite of the vast information on response to water deficits at the crop, and
molecular level, there are few examples of designed change in crop performance
through genetic improvement. This paper suggests that the lack of this knowledge
is partly due to lack of research priority focus and organization of interdisciplinary
teams. The case is made for a sequence of research activities broadly grouped under :
1) measurement of the resource and of the crop response in the components
of resource capture (transpiration), resource use (water-use efficiency) and resource
allocation (harvest index) to define targets for better focus; 2) study of mechanisms
to identify putative traits and develop an "internal consistency" in the relationship
between the traits and field performance; 3) design models at geographic, crop,
genotype (ideotype), and gene level to both priority research opportunities and to
enhance the efficiency of the improvement process; and 4) develop methods to
incorporate selection for adaptation to moisture deficit in routine plant breeding
programs.
This framework, to move from understanding to manipulation, is based on a
case study for the improvement of tropical maize ( Zea mays L.) to moisture deficits.
In tropical maize, intermittent periods of moisture stress are estimated to cause yield
losses of 15% in more than 32 million ha. Maize is particularly vulnerable around
flowering, causing a loss in grain number and through harvest index, grain yield.
Based on this understanding of the mechanism for adaptation of maize to drought,
a "drought plant ideotype" was developed. Eight cycles of recurrent selections for
the ideotype resulted in a gain of 500-800 kg/ha over a range of water-limited
environments. Yield gains were associated with increased ear growth rate around
flowering and harvest index at maturity (resource allocation).

INTRODUCTION
This paper does not review the wealth of information on rice adaptation to
variable and moisture-limiting environments. Instead it provides a conceptual
framework that will allow understanding of the physiology of adaptation to be used
effectively to manipulate the crop and the environment for enhanced productivity.
1 The International Rice Resarch Institute, P.O. Box 933, Manila, Philippines.

From the plant perspective, the crop in the field is a community of genetically,
fairly uniform plants representing the "genotype". During a growing season the
particular combination of soil, climate, and pests, plus the competition among crop
plants and other species (weeds) forms a unique environment, which when
interacting with the "genotype", leads to the expression of the phenotype, and
ultimately yield.
Much of our success in improving rice productivity was the result of enhancing
and stabilizing the environment, particularly for water (irrigation) and nutrients
(fertilizer) and at the same time increasing the genotype's ability to respond to this
change in the environment with increased yield. There was an additional benefit.
Irrigation and nutrients reduced the number of unique environments determined
mainly by temperature ranges" and the occurrence of pests. Fewer unique
environments required fewer unique genotypes to maximize the phenotypic
expression. We were able to develop rice cultivars adapted over a broad geographic
range and responsive to a given and predictable environment.
In contrast, the rainfed lowlands have a range of environments whose
diversity is driven by variability in, for example, the amount, intensity, and duration
of water. Each of the key environmental variables, eg., moisture regime, fertility,
flooding incidence, biota, daylength is complex in itself and interacts with the others
and with the rice plant. The resultant phenotype-the expression of the genotype in
these variable environments-is by way of the metabolic processes of the plant, i.e.,
its physiology. Given the environmental complexity, there are practically an infinite
number of expressions of phenotype. Thus, to advance crop production in these
environments through directed improvements in phenotype, those growth processes
that are interacting, rather than the myriad of interaction themselves must be
understood. That is, understanding leading to the rice phenotype is one means to
reduce enormous complexity to a manageable level.
Wilson (1992) contends that formal treatment of physiology gave information
as to how the crop grows, but little on how it might be manipulated genetically to
achieve particular results. This paper examines some case studies in other cereals
(but mainly maize) where such manipulations have occurred. Passioura (1981)
contends that breeding the changing of a genotype-is a practical affair with its
successes and failures being judged by the farmer clients. In contrast, the success of
physiology understanding the growth processes of the plant-depends on how
enlightening it is in providing ideas and tools for the breeding practitioners.
Our task is to develop realistic means to evaluate the contribution of
physiology research. Thus a set of "signposts" that will generate realistic ideas and
understanding, which then become appropriate tools for manipulation by the
practical breeder is proposed. These are measurements, mechanisms, models, and
methods.

SIGNPOSTS
Measurement
What to measure and why? When the new science of crop physiology began
some 40 yr ago, there was an expectation that simply understanding and measuring
such processes as photosynthesis and transpiration, would provide a direct benefit
to yield. In retrospect, we began to look at very fine-albeit critical- processes at the
very beginning without generating a broad conceptual framework of crop
development. The first measurement needed considers the basic components of the
limiting resource and how it is used, viz. resource capture, conversion and
partitioning. Under water-limiting environments, grain yield is a function of the
water transpired (T), the water-use efficiency (WUE) and the partitioning of dry
matter to grain (HI). One or more of these components must be altered to increase
grain yield:
Thus, before embarking on a breeding program for improvement under
variable and water-scarce environments, one must first measure how the current
system uses the scarce resource (water) and where the opportunities for genetic
intervention lie. We need to measure how much of the total precipitation is actually
transpired and how much is lost in runoff, soil evaporation and drainage, and how
much is stored in the soil. We must also know if there is variation in WUE and HI
(Fig. 1).

Fig. 1.

A scheme of water balance showing stored soil water (S) as a target for improvement of
sorghum under terminal drought.

Let us consider two case studies-maize and sorghum (Sorghum bicolor (L.).
In maize, Kassam et al. (1975), the pioneers of todays modelers fitted the water
requirements of maize to the bimodal rainfall patterns of the savannas in West Africa.
Their model predicted two opportunities for improving the performance of maize
in this environment (Fig. 2). In the first season, there was a high probability that
demand for water for transpiration will exceed supply midway during
growth-around flowering. Studies (Shaw, 1977; Westgate and Bassetti, 1990; Boyer,
1992) have shown flowering to be the most sensitive stage of growth in maize;
moisture stress at flowering severely reduces grain number and harvest index (HI).
Thus one target for the improvement of maize grown under variable and
unpredictable moisture supply is to modify flowering behavior and grain number
and thereby increase yield through higher HI. Ed meades et al. (1993) have estimated
that more than 32 million ha of maize grown in the lowland tropics are exposed to
intermittent drought around flowering causing 15% losses in yield.

Fig. 2

Rainfall and crop water requirement for maize, Ibadan, Nigeria.

The second case study involves improvement of sorghum in Australia, where

it is also grown in the savanna with variable and unpredictable rainfall. Maximum
yields are obtained from full-season crops but these are often exposed to a terminal
drought. In some soils, moisture remains stored at depth in the profile, even at
maturity of the crop (Williams, 1990). The target is to increase the zone of resource
capture for water in the profile and thereby the amount of water transpired (T) by
the crop and thus yield.
4

In these two examples, measurement of the limiting resources and its effect
on yield component set clear targets and focus for crop improvement. Setting such
targets is critical in the rainfed ecosystem. Because of the variability (and large
genotype environment interaction), the selected targets will not be expressed in
all locations at all times. Careful measurement allows for identifying the repeatable
and useful positive genotype environment interaction which will have greatest
benefit over time at any one location.

Mechanisms
Thew is a need to understand the mechanisms and the processes that will
increase either resoure capture, resource-use efficiency or allocation.
In the past, physiologists were content to show a correlation between a given
plant character and yield-without differentiating causality from spurious, although
statistically significant correlation. A good example of this misuse is the attempt to
improve grain yield by selecting for either an increase in grain number or grain size.
This is bound to be ineffective if grain yield is limited by the capacity to produce
assimilates for grain storage and not by the storage system of the grain itself. Not
only must the presence, strength and direction of a process be correlated with the
targeted yield component there also must be physiological "internal consistency" in
each of the intervening steps.
In sorghum, a target is to increase root exploration for soil moisture at depths
during grain filling. One mechanism for that is osmotic adjustment (OA)- the change
in osmotic potential due to the accumulation of solutes within the cell. Osmotic
adjustment maintains turgor pressure under moisture stress and thus root activity.
It also increases dehydration tolerance and in terminal stress, enhances the
retranslocation of assimilates to the grain (Santa maria et al., 1986). It is important to
demonstrate that OA is not just related to grain yield, but also that there is an internal
physiological consistency whereby
*

OA influences turgor at a given potential and influences root growth;

Increased root growth delays leaf area loss under moisture stress and
increases dry matter; and

Increased assimilate production increases yield.

Thus, the internal consistency links OA to an increase in grain yield under

moisture deficits where there is water at depth in the soil profile.
Many studies have shown maize as being unusually susceptible to drought
stress at flowering (Fig. 3) (Westgate and Bassetti, 1990; Boyer, 1992). This is partly
because male and female flowers are separated by 30-100 cm, because silk must
extend up to 30 cm to escape the husk, and because the pollination process demands
that long and delicate stigmata tissue and pollen grains remain viable even while
exposed to desiccating environments. Although a number of processes influence
grain number before, during, and after anthesis under moisture stress, the greatest
reduction occurs within the first 2 wk after pollination (Grant et al., 1989).

Fig. 3

Effect on maize grain yield from 1d of moisture stress (a) and 1d with 54% crop shading (b).

A characteristic of maize under environmental stress is increased anthesis-silkinginterval. The reserves of carbohydrate in the maize plant at silking are comparatively
small and the capacity of newly fertilized ovules to attract these reserves is limited.
Boyle et al. (1991) infused the stems with a sucrose-rich solution to offset most of
the effects of drought stress at flowering. Edmeades et al. (1993) concluded that the
reduced grain number per plant under drought seems more of a chemical (CHO
supply) than physical (separation of pollen and stigma) process, and that the
anthesis-silking interval is a symptom of reduced assimilate flux rather than the
direct cause of barrenness. Here the internal physiological consistency links the
relationship between anthesis-silking interval and yield (Fig.4), with reduced
assimilate supplies around flowering for ovule and silk growth (Fig.5) and to reduce
grain number and HI.

Fig. 4

Relationship between anthesis silking interval and grain yield combined data from maize
cultivars grown under drought (o) and well-watered conditions () at Tlaltizapan, Mexico.
From Bolanos and Edmeades (1993).

Fig. 5

Relationship between mean ear spikelet biomass at 50% anthesis and the anthesisto-silking (ASI) interval for Tuxpeno Sequia cycles 0,2,4,6 and 8 grown under high plant
density and mild severe drought stresses.

Models
Modelling processes and determining the likely effects of changes in plant
characteristics aid in the development of "best bet" approaches. Processes that cannot
be modelled are probably not understood. Without this understanding, our move
from measurement to new technologies will be unpredictable (and quite often
unsuccessful).
There are many levels of modelling-they do not have to be complex if that is
not necessary for the job at hand, The conceptual model or the "ideotype" has served
physiologists and breeders well. The most successful dates from 1960, with the IR8
story of IRRI. In 1974 Jennings, the IRRI breeder, involved in the green revolution
said:
"elegant work by the institute crop physiologists indicated in
detail on paper an ideal plant type, i.e., an "ideotype", and pinpointed
the vegetative characteristics of traditional varieties requiring drastic
modification. In retrospect it was a relatively simple task for breeders
to follow this blue print and produce a series of productive dwarfs
beginning with IR8".
Fischer et al. (1983) and later refined by Edmeades et al. (1993) developed an
ideotype for the selection of maize for better performance under limiting moisture
around flowering. They used this information in a selection index for a field site
recurrent selection program for the improvement of maize performance under
moisture deficits.
7

The ideotype concept which creates scientific ideas, and a framework within
which they can be tested should remain a strong tool for a breeding program. The
latest application is in the development of a new plant type to raise the yield frontier
in tropical rice. Originally, almost as an intuitive tool, our latest ideotypes have
involved crop modelers and physiologists from the outset. We have advanced in our
understanding of processes since the 1960s and now it is possible not only to develop
the conceptual model (as for IR8), but also process-driven models with which
sensitivity analysis-the "what if" questions- can be done in the computer before going
into the field (Kropff et al., 1994).
Models at the crop and molecular levels are now enhancing our efficiency in
research. Recent applications of process-based rice crop models linked with GIS and
to reliable data bases are allowing targeting ideotypes and assessing the performance
of alternative crop ideotypes. At the molecular level, we are developing the genetic
maps for the putative traits of root length, root penetration, osmotic adjustment and
desiccation tolerance as "models"to aid the selection process. This work is reported
in more detail elsewhere in these proceedings. And just as for past studies at the
process level, we must ensure that our molecular maps are developed from
appropriate populations to ensure that the genetic information is sufficiently robust
for application for selection within the agronomically fit population growing in the
breeder's field.

Method
This is the most critical of the signposts, requiring physiologists and breeders
to work together to be productive. The synthesis of the previous steps must come
here, or the effort invested will serve merely academic interests.
In the past, physiologists have tended to disregard genetic variation.
Extrapolation and generalization were often derived from a few, or worse, one
genotype. Correlations which depend on one extreme outrider entry were selected
from unadapted material from a germplasm bank. Too often a relationship among
yield-determining characters was established in controlled environments and never
in the field. Indeed, physiologists have worked downstream from the breeding
program spending time telling the breeders what they managed to accomplish, rather
than assisting in the current selection process.
This has led to charges that physiology is a retrospective science. The guiding
rules to applying our knowledge and skills should include (a) that physiologists
work with many genotypes representative of the variation in breders fields; and
(b) that the physiologists maintain a well-worn path to those breeders' fields.
The maize, ideotype was developed from the study of mechanisms in a broad
range of maize germplasm and from an analysis of the variation that exist within
one maize population (Tuxpeno), which is used as a breeding population for new
varieties (Fischer et al., 1983).
Using the ideotype characters for selection, divergent subpopulations were
developed and their performance compared with that of subpopulations selected
8

on yield alone. The study demonstrated that a) genetic variation for drought
characteristics existed within agronomically fit populations being used in breeding
programs; b) the ideotype selection enhanced performance under yield-limiting
moisture around flowering; c) there was an internal consistency linking change in
the population through selection to the processes that influence adaptation to
moisture deficits; d) selection for the drought ideotype was more effective than
selection for yield alone; and e) screening for the additional traits was practical (although
costly) for a breeding program.

IMPACT
Recently Edmeades et al. (1993) have reported progress after six selection
cycles for maize improvement to the target drought environment. The evaluation
studies included entries from the same breeding population where selection was for
yield (alone) done over a number of test locations (the more conventional approach
for breeding). Selection for drought tolerance resulted in a significant increase in
grain yield at levels ranging from 1 to 8 t/ha, with no significant interaction between
rate of gain and yield level. At 2 t/ha (common under drought conditions) this
genetic improvement represents a gain of 6.3% per cycle.
Gains in yield were the result of reduced barrenness under moisture dificit at
flowering. Total biomass under drought was unaffected by selection, so the gains in
grain yield arose from a systematic increase in harvest index-the target component
of selection. This was accompanied by an increase in ear biomass at 50% anthesis
and a decrease in the number of tassel branches (a change in the allocation of
resources). Flowering behavior was markedly affected by selection, with a large
decrease in the ASI, which was related to ear spikelet biomass at flowering
(Edmeades et al., 1993).
Based on this study, selection for adaptation to moisture stress around
flowering is being practiced in a number of maize populations as a routine in the
breeding program of CIMMYT.

REFERENCES
Bolanos J, Edmeades G O (1993) Eight cycles of selection for drought tolerance in lowland tropical maize.
I. Responses in grain yield, biomass, and radiation utilization. Field Crops Res. 31:233-252.
Boyer J S (1992) Mechanism for obtaining water use efficiency and drought resistance. Pages 181-200 In:
Plant breeding in the 1990's. H.T. Stalker and J.P. Murphy, eds. Commonwealth Agricultural Bureaux
International, Wallingford, United Kingdom.
Boyle M G, Boyer J S, Morgan P W (1991) Stem infusion of liquid culture medium prevents reproductive
failure of maize at low water potential. Crop Sci. 31:1246-1252.
Edmeades G O, Bolanos J, Hernandez M, Bello S (1993) Causes for silk delay in a lowland tropical maize
population. Crop Sci. 33:1029-1035.
Fischer K S, Johnson E C, Edmeades G O (1983) Breeding and selection for drought resistance in tropical
maize. El Batan Mexico.

Grant R F, Jackson B S, Kiniry J R, Arkin G F (1989) Water deficit timing effects on yield components in
maize. Agron. 1. 81:61-65.
Kassam A H, Kowal J, Dagg M, Harrison M N (1975) Maize in West Africa and its potential in Savanna
areas. World Crops 27:73-78.
Kropff M J, Cassman K G, van Laar H H (1994) Improving yield potential: the role of ecophysiological
crop models. In : Proceedings of the International Rice Research Conference, 1992 IRRI, Los Banos,
Philippines.
Passioura J R (1981) The interaction between the physiology and the breeding of wheat. Pages 191-200.
In : Wheat science-today and tomorrow. LT Evans & WJ Peacock, eds. Univ. Press, Cambridge, London.
Santamaria F J, Ludlow M M, Fukai S (1986) Drought resistance traits in hybrids and lines of Sorghum
bicolor L Moench. Pages 127-143. In: Proceedings of the first Australian Sorghum Conference, M.A. Foale
and R.G. Henzell, eds .
Shaw R H (1977) Climatic requirement: Corn and Corn Improvement. G.F. Sprague, ed. Am. Soc Agron.
18:591-623.
Westgate M E, Bassetti P (1990) Heat and drought stress in corn: what really happens to the corn at
pollination? Pages 12-28. In: Proceedings of the 45th Annual Corn and Sorghum Research Conference,
5-6 Dec 1990. D. Wilkinson, ed. American Seed Trade Association Washington, DC, USA.
Williams R L (1990) Genotypic variation in delayed leaf senescence in sorghum. MS Thesis Univerisity
of Queensland Australia.
Williams T V, Snell RS, Ellis JF (1976). Methods of measuring drought tolerance in corn. Crop Sci. 7:179-181.
Wilson G L(1992). Crop physiology: Some recollections and current perceptions. Pages 2-10 In: Proceedings
of the 6th Australian Society of Agronomy Conference, Armidale, Australia.

10

IS THERE ANAEROBIC METABOLISM IN SUBMERGED

RICE PLANTS ? A VIEW POINT
H.Greenway 1 and T.L. Setter 2

SUMMARY
Submergence of rice is a complex phenomenon acting mainly via the 104 -fold
slower diffusion of gases in solution than in the gas phase. This will lead to both
increases and decreases of supply of oxygen and CO2 and increases in the
concentration of the phytohormone ethylene. This paper focuses on possible O2
deficiencies; without claiming that O2 deficiency is more important than other
possible adverse effects of submergence. The occurrence of O 2 deficiency in
submerged rice is reasonably certain. However, we do not know whether such O2
deficiency will affect survival, or at least long-term growth and yield of the crop.
Low O2 concentration occurs in the water of flooded rice fields, particularly
at greater water depths and also during the night. Available evidence suggests that
O2 deficiency occurs, in partially or wholly submerged rice plants, at least at certain
hours of the day and in certain regions. During the night, O2 concentration at the
epidermis of the root tips is zero and the roots synthesize ethanol, the end product
of anaerobic catabolism. All this ethanol is consumed during the first hours of the
light period. This suggests the possibility that ethanol may be used as a source of
carbon, both in submerged rice and in nonsubmerged plants of paddy fields.
Armstrong's model indicates that O2 deficiency in roots of vascular plants is
particularly likely in the stele of the root tips; O2 supply from the environment to
this core being limited by diffusion and O2 uptake along the pathway. We review
metabolic evidence for roots which is consistent with these models, no information
being available on submerged leaves.
In reviewing mechanisms of anoxia tolerance, we assume universal
requirements for retention of membrane integrity despite a greatly reduced
energy supply and reductions in requirements of energy for maintenance. Using
these requirements as a prerequisite, we present evidence that there are two
different modes of adaptation to anoxia, based on (i) rapid and (ii) slow rates of
catabolism, respectively. We also suggest anoxia tolerant plants with the mode
of slow rates of catabolism may be particularly useful during transient O 2
deficiency as often occurs in floodprone rainfed environments.

1 The University of Western Australia, Nedlands. West. Australia

2 The International Rice Research Institute,

P.O. Box 933, Manila, Philippines

Injury due to exposure to anoxia can also be aggravated upon return from
anoxia to aerated conditions, for example production of free radicals of oxygen
would aggravate membrane damage and injury in several species. Whether this
occurs in rice is unknown.
The response to anoxia, which involves a large-reduction in energy production
is also relevant to other situations, for example to carbohydrate starvation, which
can in principle occur due to other factors of a submerged environment such as low
CO2 supply and low light.

INTRODUCTION
The complex environment in flooded rice fields is shown in Fig. 1. The 104 -fold
slower diffusion of gases in solution than in the gas phase (Armstrong, 1979) means
the solution immediately outside the shoot epidermis can contain high O2
concentrations during the day and low concentrations during the night. Theopposite
holds for CO2 concentrations. This conclusion receives support for O 2 from data by
Setter et al. (1988a) for gas concentrations in floodwater in rice fields in Thailand.
These studies also showed that floodwater contained high concentrations of the
phytohormone ethylene (Setter et al., 1988b).
The rest of this review will be confined to the posssible occurrence of O2
deficiency in submerged rice and if so what mechanisms of adaptation may be
required. However, it should be pointed out that one requirement for tolerance to
anoxia, ability to survive at a low level of energy production, is also relevant to cases

Fig. 1.

12

Diagram showing the large range of adverse factors which may occur during submergence
of rice (Setter et al., 1993).

of carbohydrate starvation, whether these are due to low CO2 low light, or chlorosis
caused by high ethylene concentrations, which develops in a submergence intolerant
cultivar (Jackson et al., 1987).
Occurrence of O 2 deficiency depends on :
1)

Plant factors, reviewed by Armstrong and Beckett (1987) for mots in

waterlogged soils depending on O2 supply from the shoots. These
include: i) distance from the O 2 supply, longitudinally via the
aerenchyma, as well as radially from the aerenchyma through the cortex
into the stele, ii) porosity of the tissues, iii) rate of O 2 consumption in
respiration and iv) radial O2 loss to the environment. The integration
of these four factors will favour the earliest development of anoxia in
the stele of the root tip.

2)
i)

Environmental factors as discussed by Setter et al. (1988a), being:

The degree of turbulence in the water, which will determine the O2
movement to the tissues, and
The temperature, which will affect O 2 consumption by both plants and
micro-organisms.
O2 production during photosynthesis. Hence O2 supply to submerged
rice will vary with the diurnal cycle and also depend on light intensity,
as affected by shading and turbidity, and on CO2 concentration, which
will be a function of total inorganic carbon and the pH of the floodwater.

ii)
iii)

The transient nature of any O2 deficiency in submerged rice means that we

can not deduce from the literature whether O 2 deficiency contributes to the adverse
effects of partial or complete submergence.
In physiological terms we can only speculate on the best mode of adaptation.
Adaptation by plant cells to O2 deficiency is complex and includes the capacity to
sustain and regulate alcoholic fermentation ability to reduce maintenance
requirements for energy and maintenance of membrane integrity. We present
evidence for different rates of metabolic activity in various tissues.
1)

Low metabolic activity coupled with survival, i.e. by the tissue going
into suspended animation, thus conserving carbohydrates for recovery
after the floodwaters recede.

2)

High metabolic activity, sometimes coupled with partial maintenance

of growth and nutrient uptake.

Possible "post anoxic injury" due to return to aerated conditions has also to
be considered. Transfer of tissues from anoxia to air, i.e. restoring the oxygen supply.
Introduction of O 2 may lead to formation of free radicals of oxygen and further
membrane damage, particularly when there has been deterioration of membrane
components during the O2 deficiency itself.
Finally, the relevance of the present paper is for the rice ecosystem rather than
for rice per se : the rice weed, barnyard grass, is probably even more efficient than
rice growing under anoxic conditions. Understanding the mechanisms of the
13

adaptation of barnyard grass could lead to more efficient methods of its control
(Kennedy, et al., 1993).
Definition of Terms
anaerobic catabolism : breakdown of sugars to supply energy in the absence of air.
anaerobic proteins: proteins induced by O2 deficiency.
anoxia: absence of oxygen.
anoxic core: centre of a tissue which does not receive oxygen.
ATP: adenosine trios-phosphate.
daltons: a measure of molecular size similar in magnitude to molecular weight.
glycolysis: breakdown of sugars to the three carbon compound pyruvate.
hypoxia: low but not zero oxygen.
Km: substrate concentration at which half the maximum velocity of a reaction is
attained.
maintenance requirement: energy required without any increase in cell substances.
NAD: nicotinamide adenine dinucleotide.
O2 deficiency: Plant tissues can not obtain sufficient oxygen for maximum respiration
post anoxia: after return from anoxia to air.
PPi : pyrophosphate.
RNA: ribonucleic acid.

DETAILED REVIEW
Occurrence of O2 deficiency in the field
Low O2 concentrations have been shown at greater depth and during the
night in floodwaters of fields of floating rice in Thailand (Setter et al., 1987, 1988a)
and India (Setter et al., this proceedings). O2 concentrations in the floodwaters
decreased with water depth, while in the surface layers of water O2 concentrations
increased during the day and decreased during the night. O2 concentrations inside
floating rice plants are only available for the lacunae, and these show the expected
decrease in O2 concentrations with increasing distance from the O2 supply, i.e. with
increasing water depth (Setter et al., 1987). Similar patterns of O2 distribution in roots
of paddy rice were established in the pioneering studies by van Raalte (1940) in the
Botanical gardens of Bogor. Such patterns are predicted by the elegant models on
O2 distribution within mots of vascular plants, depending on O2 supply via the
aerenchyma (Armstrong and Beckctt, 1987).
No such models for O2 supply exist yet for submerged plants, nevertheless
we can use the model of Armstrong and Beckett to predict that : (i) during the night,
14

O2 supply to roots of submerged rice plants will be substantially lower for submerged
than for non-submerged plants: with O2 close to the shoot root junction at 7% for
the submerged plants (Setter et al., 1987) and an expected 20% O2 in the gas spaces
of non-submerged plants(Armstrong and Beckett, 1983, (ii) O2 supply is likely to
be less for tissue away from the principal O2 supply, i.e. the lacunae or aerenchyma.
As one example, in deep water rice O2 deficiency may become acute in aquatic roots
in deep water unless their epidermis has a low permeability to O2 .
Different modes of response by vascular plants to O2 deficiency
We distinguish four different modes of response of plants to low O 2
concentrations in at least part of their environment.
Mode 1 : intolerant to waterlogging or submergence. These plants have both low porosity
and low cellular tolerance to anoxia and comprise the bulk of plant species.
Mode 2: tolerant to waterlogging and submergence based on high porosity of tissues
to facilitate O2 flow from the air or from water containing O2 (Armstrong, 1979, Armstrong
and Beckett, 1987). Some plants with this mode are not necessarily tolerant to anoxia (cf.ap
Rees et al., 1987) and this has been claimed for rice roots (Webb and Armstrong, 1983). We
do not agree and include rice into Mode 4, as discussed below.
Mode 3: tolerance based on cellular tolerance to anoxia. This mode can be further
divided into two types: 3A) where a tissue can grow under anoxia, for example, rice coleoptiles
(Jackson and Drew, 1984; see also Fig.2) and some species of Echinochloa (Kennedy et al.,
1992), 3B) tissue survives, but does not grow until aeration is restored. The best examples
are germinating seeds of several species (to be discussed in section 3).

Fig. 2.

Genotypic differences in anoxia tolerance of rice during germination (Setter et al., 1994).
Conditions : seeds imbibed 1 day air, followed by 3d anoxia or air treatment.

15

Mode 4: a tolerant mode based on a combination of Modes 2 and 3. This mode can
cope with temporal decreases in O2 regime for example during diuranal cycles of O2 supply.
This combined mode also allows for "symbiosis" between anoxic and aerobic regions in the
same tissue. Well known examples are aerenchymous root tissues with anoxic steles and
aerobic cortices (Armstrong and Beckett, 1987). Submerged established rice plants
presumably belong in this group (to be discussed in more detail in Section 5).

Anoxia tolerance and the production and flow of energy under anoxia
We postulate that universal requirements for anoxia tolerance are the
reduction of maintenance requirements for energy and the maintenance of
membrane integrity, Evidence for the second proposition was obtained for beet root
in which the endogenous dye betacyanin is an excellent marker for cell death (Zhang
et al., 1992). Rapid leakage of K+ , C1- and Na + started up to 9h before leakage of
betacyanin indicating increases in membrane permeability caused death, rather than
vice versa (Zhang et al., 1992, Zhang and Greenway, 1993).
Production and flow of energy to processes essential to survival. Generation of
energy during anoxia will occur mainly via glycolysis {equation 1) with ethanol as
the principal end product of anaerobic catabolism in nearly all plants (equation 2 :
Davies, 1980; ap Rees et al., 1987). The overall reactions are:

The purpose of ethanol synthesis (equation 2) is to enable energy production

to continue from glycolysis(equation 1) by recycling the essential cofactor NAD;
these combined pathways are referred to as alcoholic fermentation. The production
of energy from alcoholic fermentation per mol hexose is only 2 mol of ATP, compared
with 24-36 mol ATP produced in aerated tissues mainly by oxidative
phosphorylation. The energy yield under anoxia may be increased to 2.5 mol ATP
per mol hexose units when these units are produced from sucrose via the enzyme
sucrose synthase(based on Stitt , 1990). In some tissues the large reduction in energy
production under anoxia can be alleviated by increases in rate of sugar consumption,
traditionally referred to as the Pasteur Effect (Drew, 1990). This effect is expressed
as the ratio of carbohydrate flux through glycolysis under anoxia over the rate of
glycolysis in air (Table 1, the ratio is called from now on Canoxia/Cair). The
determination of this ratio allows distinction of two mechanisms of adaptation to
anoxia such as 1) rapid catabolism and 2) slow catabolism.
1)

16

Rapid catabolism. indicated by high ratios of C anoxia / Cair found in

rice coleoptiles and aged storage beetroot and carrot roots during
the first 24 h of anoxia (Table 1). None of these increases in rates of
sugar consumption come close to compensating entirely for the
inefficiency of energy production during glycolysis i.e. 14 to 18-fold
increases in carbohydrate flux would be required for equivalent
energy production (calculated from above).

2)

Slow catabolism: indicated by low ratio of Canoxia/Cair is found for

storage tissue after prolonged exposure to anoxia. More spectacularly,
germinating lettuce seeds exposed to anoxia went into suspended
animation (Table 1). Energy production in these seeds must have been
negligible, yet these seeds germinated when they were returned after
14 days anoxia to air(Pradet and Bomsel, 1978). Similar pattern have
been found for an Echinochloa crus pavonis (Zhang et al., 1994).

Table 1.

Different rates of catabolism in anoxia tolerant species. The change in

rate of glycolysis is expressed as the ratio for carbon flow through
glycolysis: Carbon anoxia / Carbon air .

Plant

Carbonanoxia /
Carbon air

Reference

1.8

Calculated by us from
Alpi and Beevers (1983)

Carrot storage tissue

(first h of anoxia)

3.0-3.7

Faiz-ur Rahman et al.

(1974)

Beetroot storage tissue

(first 2A h anoxia)

1.9

Zhang and Creenway

(1994)

<1.0

Pradet and Bomsel(1978)

Raymond and Pradet(1980)

0.7

Zhang and Greenway(1994)

A) High rates of catabolism

Rice seedlings

B) Low rate of catabolism

Seeds of several spp.
Aged beetroot tissues
(>2A h anoxia)

The energy produced during glycolysis in anoxic tissues is used for different
processes:
1)

Maintainance. This would include turnover of proteins and membrane

components and essential solute transport, e.g.sugar and proton
transport. This reqirement presumably exists in most tissues excepting
the seeds which go into suspended animation as described above.

2)

Synthesis of anaerobic proteins which, mainly function in sugar catabolism

(for maize, Bailey Serres et al., 1988) and perhaps synthesis of specific
membrane component of adaptive value under anoxia. The synthesis
of anaerobic proteins is particularly pronounced during the first 24 h
of anoxia; and may explain the higher rates of catabolism during the
first 24 h of anoxia than after 24 h in the beetroot tissues (cf. Table 1).

17

3)

Growth. Elongating rice coleoptiles require net cell wall and protein
synthesis (Alpi and Beevers, 1983), as well as uptake of solutes to
generate osmotic pressure (Atwell et al., 1982). The high energy
requirements, particularly for protein synthesis, may account for the
high rate of catabolism found in rice coleoptiles (Table 1).

The above requirements for energy have to be considered in relation to the

large reduction in rates of ATP synthesis discussed earlier; even the high rate of
catabolism in the carrot tissue (Table 1) indicates ATP yields under anoxia are 3-6
fold lower than that in air. So in all anoxia tolerant species there must be a
sophisticated regulation of energy flow. One type of regulation is a reduction in
maintenance requirements for energy. Such requirements can be best assessed for
storage tissue which does not grow even under aerated conditions. For aged tissues
of storage root of red beet we have estimated that the maintenance requirement for
energy is 10-25 times lower under anoxia than under air (Zhang and Greenway
1994). This reduction is so large that we assume there is reduced turnover of protein
and to a lesser extent membrane components since these processes use the bulk of
the energy for maintenance in aerobic cells (Penning de Vries, 1975).
Examples of regulation. Examples of regulation of various processes under
anoxia can be found at different levels of organisation.

18

1)

Echinochloa crus pavonis investigated by Zhang et al. (1984) can: a) either

survive for at least 30 days in suspended animation (anoxia from start
of imbibition), or b) develop a coleoptile and hypocotyl after aerobic
treatment for the first 1-18 h followed by transfer to anoxia.

2)

Regulation is demonstrated both by the switching on of synthesis of

anaerobic proteins and the cessation of synthesis of many other proteins,
even though their messenger RNA still persists (Drew, 1990). For example
in rice cultured cell there are substantial increases in activities of several
enzymes (Table 2) including PPi dependent phosphofructokinase. In
contrast, ATP dependent phosphofructokinase changed little in activity,
presumably because activity was already high under aerated
conditions.

3)

There are much larger reductions in transport of Cl than of sugar even

in anoxia tolerant rice coleoptiles (I. Zhang and II. Greenway, 1994.
person. commun.). A reasonable teleological speculation is that the cells
direct limited amounts of energy to the essential transport of sugars
required for anaerobic catabolism, rather than to Cl. However, we do
not know by what molecular mechanism this regulation is achieved.

4)

Rice shoots achieve a remarkable regulation of the pH of the cytoplasm.

After sudden transfer from 20% to zero O2 the pH(cyt) decreased first
to 7.0 but recovered to 7.15 after 10 h at 1%, O2 compared with a pH(cyt)
of 7.4 to 7.5 in aerated shoots (Menegus et al., 1991). In this case there
is a plausible mechanism of regulation of energy flow since the tonoplast
of plant cells contains a PPi dependent-H + translocase in addition to

the more ubiquitous ATP dependent-H+ translocase (Rea and Poole,

1993). Hence conversion of ATP to PPi might be one way to direct a
limited amount of energy to the crucial process of preventing acidosis
of the cytoplasm.
Table 2.

Increases in activity of several enzymes under anoxia indicated

restricted synthesis of a set of proteins, the anaerobic proteins, most of
which are involved in sugar mobilisation and catabolism. Data for
cultured rice cells by Mohanty et al. (1993).

Enzyme

increase over activity found in air

Alcohol dehydrogenase

7 fold

Lactate dehydrogenase

4 fold

Sucrose synthase

3 fold

PPI

7 fold

dependent-phosphofructokinase

Importance of hypoxic pretreatment (acclimatizing) to anoxia

Anoxic shock often leads to rapid injury which can be prevented in some
tissues by acclimatizing them by exposure to low but not zero O2 concentrations
before exposure to anoxia (reviewed by Drew, 1990). As one example, survival of
root tips of intact maize plants is extended from less than 24 h anoxia to 96 h anoxia
by acclimatizing (Drew, 1990). In many soils it is likely that O2 concentrations
decrease gradually when soil become waterlogged (Drew, 1990), emphasising the
relevance of data obtained using hypoxic pretreatment. Furthermore when data are
obtained using anoxic shock it remains uncertain whether the changes in metabolism
measured are merely a consequence of injury and approaching death or a more
subtle metabolic response to anoxia.
One reason for the protective value of hypoxic pretreatment is the observed
synthesis of ADH and therefore presumably also other anaerobic proteins (Johnson
et al., 1989), while the tissue as a whole has still substantial oxidative phosphorylation
(Johnson et al., 1989). Thus once the tissues are transferred to anoxia they would face
smaller demands on the restricted energy production than tissues exposed to anoxic
shock (Zhang and Greenway, 1994) . Alternatively ADH may not be synthesised at
all following anoxic shock (Johnson et al., 1989).
Evidence for coexistence of anoxic and aerobic tissues in the same organ
This section deals with the central issue of this paper since it is unlikely
that submerged established rice plants are ever completely anoxic. This contrasts
with the early developmental phase during direct seeding when the whole
19

seedling may become anoxic. Once O2 supply to a tissue becomes restricted rather
than cut off altogether anoxic and aerobic tissues will coexist in the same organ.
The hypothesis that this combination occurs is based on the very high affinity of
cytochrome oxidase for O2 indicated by a Km of 24 m molm-3 (Yocum and Hackett,
1957). So electron flow and oxidative phosphorylation will continue till very low
O2 concentrations are reached. As shown schematically in Fig. 3, O2 will diffuse
into the tissue, while O 2 consumption will decrease the O2 concentration along
the pathway. In environments of low external O2 concentration the concentration
gradient will be insufficient to deliver O 2 to the core of the root i.e. this core will
become anoxic. Analogous arguments hold for other tissues, such as submerged
leaves and when the O2 supply is the aerenchyma rather than the external
environment.

Fig. 3.

Diagram showing the O 2 profile expected in a root when the O 2 supply is from the nutrient
solution surrounding the roots (adapted from Armstrong and Beckett, 1987).

Evidence for anoxic steles and aerobic cortices in maize mots.

Anoxic cores in aerenchymous rice mots have been demonstrated using an O 2 micro
electrode (Fig. 4). The radial pattern of O2 concentrations is consistent with earlier
formulated sophisticated models (Armstrong and Beckett, 1987) based on : 1)
porosity of the tissues; 2) O 2 uptake by respiration; 3) radial O 2 loss to the
environment and 4) length of longitudinal pathway. No such models exist for
submerged leaf tissue but these could be developed in cooperation between
biologists and mathematicians.
The data on O2 concentrations are complimented with metabolic evidence
for an anoxic stele and aerobic cortex in maize roots (Thomson and Greenway,
1991), exposed to low but not zero O2 concentrations or depending on O2 supply
20

from the shoots via the aerenchyma. Evidences are as follows:

1)

Higher concentrations of ethanol in the stele than in the cortex (Table 3).

2)

An increase in ratios of alanine to free amino acids in the stele but

not in the cortex of hypoxic roots (alanine is an end product of
anaerobic catabolism: ap Rees et al., 1987).

3)

"Active" states of the key enzyme pyruvate decarboxylase (PDC) in the

stele of hypoxic roots while the same roots have "inactive" states of this
enzyme in the cortex (Table 4). This is key evidence since PDC is the
entry port for carbon flow to ethanol and hence will determine whether
carbon of glycolysis flows to ethanol or into the aerobic catabolic
pathway of the Krebs cycle.

Table 3.

Ethanol (mol m-3 ) in cortex and stele of excised maize mots exposed to
different O2 concentrations at 25C. (Thomson and Greenway, 1991).

O2 external (mol m-3 )

stele

Ethanol (mol m-3 )

cortex

external solution

0.06*

0.39

0.12

0.11

0.00

0.34

0.43

0.42

Air saturated water at 25C contains 0.26 mol m-3 O2 .

In Table 4, the inactive and active state of PDC in the cortex and stele,
respectively, is shown by: (i) a lag after addition of pyruvate i.e. the enzyme is
gradually activated in the presence of its substrate. In hypoxic roots such lags are
pronounced in the cortex but slight in the stele, ii) differences in activation of the
isolated enzyme when incubated at pH 6.0; these were large for the cortex but small
for the stele of hypoxic roots (Table 4) suggesting most of the PDC is already
activated.
Evidence for occurrence of anoxic regions in submerged, established rice plants :
The following evidence is available :
1)

O 2 concentrations measured with a cylindrical O2 micro electrode

surrounding the tip of nodal roots of submerged rice plants
demonstrated that O2 becomes very low during dark periods, and
increases substantially during light periods (Fig.5); this increase was
associated with O2 evolution in the shoots during photosynthesis
(Waters et al., 1989). Concurrently root elongation commenced a few h
after the start of the light period and ceased at the start of the dark
period (Fig. 4).
21

Fig. 4.

22

O 2 concentrations (A) and growth of root tips (B) during submergence of rice. O2 was
measured with a micro O 2 electrode at the epidermis of the 2-7 mm zone behind the root
apex of submerged rice plants (Waters et al., 1989). Solid bars on the x-axis represent dark
periods.

Table 4.

Active and inactive states of pyruvate decarboxylase to diagnose

blockage of oxidative phosphorylations in cortex and stele of maize
roots at different O2 regimes (from Thomson and Greenway, 1991). The
lag indicates how fast the maximum rate is established upon addition
of substrate (pyruvate); when the enzyme is fully active the lag will be
only a few seconds. The % increase in activity shows the increase in
activity after incubating extracts for 1-2 h at pH 6.0 at 22-25C.

Cortex

% increase in activity
Stele
Cortex

1.1

1.9

63

68

0.0 mol/ m -3 roots,

shoots in air

0.2

1.4

25

50

0.0 mol/ m -3 shoots

and roots

0.3

0.7

20

23

O2 Treatment

Stele

0.26 mol/ m-3 *roots,

shoots in air

Lag (min)

* Air saturated

2)

Roots of intact submerged rice plants produced ethanol during the night
and consumed ethanol at the start of the light period (Fig. 5). Mirror image
diurnal cycles were found for O2 with loss from the roots during the day
and O2 uptake during the night (Fig. 5)

The demonstration that anaerobic metabolism occurs at least in roots of

submerged intact rice plants emphasises the importance of observed genotypic
differences in anoxia tolerance in germinating rice (Fig. 2). No information is
available for established plants. However, the two most anoxia tolerant genotypes,
Kurkaruppan and FR13A, were selected during submergence of established
seedlings. So, there is strong inference that anoxia tolerance contributes to the ability
of established rice seedlings to survive submergence.
Available information on ethanol consumption by plant tissues
The evidence for coexistence of aerobic and anoxic metabolism in different
regions of the same organ opens the possibility of metabolic exchange between these
regions. This exchange is in principle possible since transport through the
plasmodesmata has been suggested for solutes with molecular weights between 465
and 1830 Daltons (Robads and Lucas, 1990).ATP has a molecular size of 507 Daltons
and can pass through gap junctions between animal cells (Darnell et al., 1990);
plasmodesmata are analogous to the gap junctions in animal cells. We can therefore
envisage flow of end products of anaerobic catabolism from anoxic to aerobic regions
23

Fig. 5.

Ethanol synthesis during night and ethanol consumption during the day by roots of intact
submerged rice plants (Waters et al., 1989). The roots were in a sealed container so that
O2 and ethanol in the nutrient solution surrounding the roots could be measured. Solid bars
on the x-axis represent dark periods.

where they can contribute to the carbon balance. Similarly, compounds able to
contribute to energy supply such as ATP and PPi may be transported from aerobic
to anoxic regions. Presumably such an input would allow much more synthetic
activity than when the entire tissue was anoxic. Most of this argument is mere
speculation and the only evidence available is for ethanol shown in Fig 5 and by
transport of ethanol from waterlogged roots of tomato plants to their shoots (Fulton
and Erickson, 1964).
In this paper we further consider possible ethanol metabolism in the aerobic
regions of an hypoxic tissue without claiming this is the only or most important
exchange of metabolites occurring in tissues with aembic and anoxic regions. Rates
of ethanol consumption are of the same order of magnitude as those of ethanol
formation (Table 5). The observed rate of ethanol consumption by rice mots is 4-fold
lower than the maximum observed rate in peas (Table 5), however, this could be
due to the difference in exogenous ethanol concentration rather than due to species
difference.
During ethanol consumption, acetate is formed together with 2 mols of
NADH/mol of ethanol as observed for pea and mungbean tissues (Cossins, 1978).
The carbon is subsequently incorporated in a large range of metabolites, similar to
carbon derived fmm glycolysis (see Cossins, 1978 for review). The pathway from
ethanol to acetate can continue as long as the tissues are aerated because NAD can
then be regenerated from NADH in several reactions and by conversion to ATP in
oxidative phosphorylation.
24

Table 5.

Comparisons between rates of ethanol consumption and

formation. All data are in m mol g -1 fresh weight -1 min-1.

Ethanol consumption in rice roots (0.33 mol/mm-3

external ethanol: calculated from Fig 5)

0.033

Ethanol consumption for peas (maximum observed;

150 mol m -3 external ethanol, Cossins and Beevers, 1963)

0.12

Ethanol formation in rice roots(John and Greenway, 1976)

0.042

ethanol

We have also substantial information on the characteristics of alcohol

dehydrogenase. The Km for ethanol conversion to acetaldehyde is 10 mol m-3 for
rice roots (John and Greenway, 1976) i.e. the enzyme had a low affinity for ethanol.
However, a substantial range in Km for ethanol between 35 and 34 mol m-3 was
found for isoenzymes of barley (Mayne and Lea, 1985) indicating the potential for
enzymes with a lower Km i.e. with a high affinity for ethanol and therefore more
suitable to ethanol consumption at low endogenous or exogenous levels. The in vitro
activity of ADH for ethanol consumption indicated that there is 12 times more activity
than the observed in vivo rates of net ethanol synthesis (Table 5). This assessment of
in vitro activity allows for pH of the cytoplasm, the known pH spectrum of the
reaction, a cytoplasmic pH of 7.5 and assumes an exogenous substrate concentration
of 0.33 (calculated from Waters et al., 1989). The present data indicate the potential
for ethanol as a carbon source and this also applies for any ethanol produced in the
soil. That is ethanol and other organic solutes occurring in waterlogged soils may
supplement carbon gain derived from photosynthesis. These alternate carbon
sources may be particularly important during periods of low potential
photosynthesis i.e. at low light intensities. There is at present no evidence available
to support or reject this hypothesis which in our view merits ecological and
physiological research. Further elucidation of the possible importance of ethanol
consumption in aerobic regions of hypoxic tissues of rice requires the following
information:
1)

Concentrations of ethanol in soil and in plant tissues.

2)

Rates of ethanol
concentration.

3)

Consumption of internal ethanol with time sequences so as to evaluate

the relation between endogenous substrate and rate of reaction in vivo.

4)

Comparison between genotypes.

5)

Comparisons between different types of tissues.

6)

Comparison of in vivo data with activities of ADH at the same substrate

concentration and pH as found in tho cell.

consumption

as

related

to

external

ethanol

25

7)

Fate of ethanol using "realistic" external concentrations rather than the

very high concentrations used in previous studies.

Possible post anoxic injury

Some of the adverse effects of exposure to anoxia may occur following return from
anoxia to air rather than during exposure to anoxia. Specifically partial break down
of fatty acids of membranes during anoxia may lead to formation of free radicals of
oxygen when this becomes available leading to further damage to membranes.
Strong evidence for this possibility- has been presented by Monk et al. (1987),
Crawford (1993), Crawford and Wollenweber-Ratzer (1992). Data for rhizomes of
marsh species are summarised in Table 6. The anoxia tolerant species had less lipid
degradation products and generated less free oxygen than the anoxia intolerant
species (Table 6).
Table 6.

Evidence for formation of free radicals of O 2 in rhizomes of an anoxia

tolerant species, Iris pseudacorus, and anoxia intolerant species,
I.
germanica and Glyceria maxima. No evidence for free radical formation
was found in the anoxia tolerant species I. pseudacorus (Crawford, 1993.
Monk et al., 1987, 1989).

Species

Lipid degradation
products (1)

Free radical
formation (2)

Superoxide
dismutase(3)

Iris pseudacorus
(tolerant)

Few

None observed

Yes (13x increase)

Iris germanica

Large number

Rapid genera-

Either decreased

tion

or small increase

Glyceria maxima
(intolerant)

Furthermore, the anoxia tolerant species increased 13 fold in superoxide

dismutase, an enzyme, which contributes to removal of free radicals:
(3)
The synthesis of the enzyme superoxide dismutase in the anoxia tolerant
species occurs under anoxia, i.e. it is one of the anaerobic proteins, yet it becomes
only functional after return to air. That some of the limited amount of energy
produced under anoxia is directed to the synthesis of an enzyme required only after
relief of anoxia is an example of the remarkable subtlety of adaptation of plants to
adverse environments. Further evidence for the existence of free radicals of O 2
consists of the improved health of chickpeas when treated with ascorbic acid, a free
radical scavenger, before returning seedlings from anoxia to air (Fig.6). One key
26

Fig. 6.

Ameliorating effect of addition of a scavenger of oxygen radicals, ascorbic acid at 10 mol

m-3 in the root medium (pH 5.6) on post anoxic recovery as evaluated by shoot length of
chickpeas after exposure to anoxia for 24 or 72 h followed by exposure to air for 7 days :
shoot branch length was measured after 7d aeration. Ascorbate was in distilled water. Control
plants were aerated continuously (from Crawford and Wollenweber Ratzer, 1992).

question is whether the failure of anoxia intolerant species to synthesise superoxide

dismutase during anoxia is a primary factor or a consequence of injury. The latter
possibility is indicated by the much more pronounced post anoxic injury after 1 h
than after 5 h anoxia, concurrently, free radical generation was higher and levels of
superoxide dismutase were lower after 1 h than after 5 h anoxia (Van Toai and Bolles,
1991).

CONCLUSION
Anoxia tolerance
We postulate two modes of anoxia tolerance. Both modes require maintenance
of membrane integrity and sophisticated regulation to direct the limited amounts of
energy to processes essential to survival, however, they differ in the rate of
catabolism :
27

i)
ii)

high rates of anaerobic catabolism

low rates of anaerobic catabolism despite a high sugar supply

Occurrence of anoxia in submerged rice plants

i)
ii)

iii)
iv)

Long term anoxia of all plant tissues probably only occurs under
flooded conditions during the early period after direct seeding.
Anoxic regions coexisting with other regions receiving sufficient O2 for
oxidative phosphorylation may be the norm for submerged rice and
may even occur in the roots of paddy rice. At least in submerged rice
the anoxic regions will be more voluminous during the night than
during the day.
The best mode of anoxia tolerance for the association of aerobic and
anoxic tissue is unknown.
It is also unknown whether O 2 deficiency adversely affects submerged
rice plants.

General comment on research approach

Clearly, adaptation to submergence is a complex phenomenon and there are
no easy solutions. Scientists dedicated to increasing plant productivity may well
question whether research on elucidating the mechanisms of adaptation to
submergence is worthwhile. This question is common to many studies on adverse
environments; one has to choose between pragmatic research, elucidation of
mechanisms of adaptation, or a combination of both. The proportion of the different
approaches depends on the scale of the problem.

REFERENCES
Alpi A, Beevers H (1983) Effects of O2 concentration on rice seedlings. Plant Physiology 71:30-34.
ap Rees T, Jenkin L E T, Smith AM, Wilson PM (1987) The metabolism of flood tolerant plants. Pages
227-238. In : Plant life in aquatic and amphibious habitats. R.M.M. Crawford.ed. Blackwell Scientific
Publication, Oxford.
Armstrong W (1979) Aeration in higher plants. Pages 226-332 In: Advances in Botanical Research. H.W.W.
Woolhouse, ed. Academic press, London.
Armstrong W, Beckett PM (1987) Internal aeration and the development of stelar anoxia in submerged
roots. New phytologist 105:221-245.
Armstrong W, Cringle S, Brown M, Greenway H (1993) In: Interacting stresses on plants in a changing
climate, NATO advanced workshop AS1 Series 1 Vol 16. M.B. Jackson and C.R. Black, eds. Wye College.
Atwell BJ, Waters I, Greenway H (1982) The effect of oxygen and turbulence on elongation of coleoptiles
of submergence tolerant and submergence intolerant rice cultivars. J.Expt. Rot 33: 1030-1044.
Bailey-Serres J, Kloeckener Gruissem B, Freeling M (1988) Genetic and molecular approaches to the study
of the anaerobic response and tissue specific gene expression in maize. Plant Cell and Environment 11:
352-357.
Cossins E A (1978) Ethanol metabolism. Pages 169-202 In: Plant life in anaerobic environments. D.D. Hook
and R.M.M. Crawford, eds. Ann Arbor Science Publisher Inc, Michigan.

28

Cossins E A, Beevers H (1963) Ethanol metabolism in plant tissues. Plant Physiology 38: 375-380.
Crawfod R M M (1993) Plant survival without Oxygen. Biologist 40: 110-114.
Crawford R M M, Wollenweber-Ratzer B (1992) Influence of L- Ascorbic Acid on Post-anoxic growth and
survival of chickpea seedlings (Cicer arietinum L.) J.Exy. Bot. 43: 703-708.
Darnell J, Lodish H, Baltimore D (1990) Molecular Cell Biology 2nd Edition Scientific American Books
W.H. Freeman and Co. New York.
Davies D D (1980) Anaerobic metabolism and the production of organic acids. Pages 581-611. In: The
Biochemistry of Plants. Vol.2 D.D. Davies. ed. Academic Press, London.
Drew M C (1990) Sensing soil oxygen. Plant Cell and Environment 13: 681-693.
Faiz-ur-Rahman A T M, Trewavas A J, Davies DD (1974) The Pasteur effect in carrot root tissues. Planta
118 : 195-210.
Fulton J M, Erickson A E (1964) Relation between soil aeration and ethyl alcohol accumulation in xylem
exudate of tomatoes. Proc. Soil Sci. Soc of America 28 : 610-614.
Jackson M B, Drew MC (1984) Effects of flooding on growth and metabolism of herbaceous plants. Pages
47-128 In:Flooding and plant growth. T.T. Kozlowski, ed. Academic Press, London.
Jackson M B, Waters I, Setter TL, Greenway H (1987) Injury to rice plants caused by complete submergence:
a contribution by ethylene(Ethene). J.Exp. Bot. 38:1826-1838.
John C D, Greenway H (1976) Alcoholic fermentation and activity of Some enzymes in rice roots under
anaerobiosis. Aust J. Plant Physiol. 3:325-336.
Johnson J, Cobb BG, Drew M C (1989) Hypoxic induction of anoxia tolerance in root tips of Zea maize. Plant
physiology 91:837-841.
Kennedy R A, Rumpho M E, Fox TC (1992) Anaerobic metabolism in plank. Plant Physiology. 100:1-6.
Mayne R G, Lea P J (1985) Properties of three sets of isoenzyme of alcoholdehydrogenase isolated from
barley (Hordeum vulgare). Phytochemistry 24: 1433-1438.
Menegus F, Cattaruzza L, Mattana M, Beffagna N, Ragg E(1991) Response to anoxia in rice and wheat
seedlings. Plant Physiology 95: 760-767.
Mocquot B, Prat C, Mouches C, Pradet A (1981) Effect of anoxia on energy charge and protein synthesis
in rice embryo. Plant Physiology 68:636-640.
Mohanty B, Wilson P M, ap Rees T (1993) Effects of anoxia on growth and carbohydrate metabolism in
suspension cultures of soybean and rice. Phytochemistry 34: 75-82.
Monk L S, Fagerstedt K V, Crawford R M M (1987). Superoxide dismutase as an anaerobic
polypeptide. Plant Physiology 85: 1016-1020.
Monk L S, Fagerstedt K V, Crawford R M M (1989) Oxygen toxicity and superoxide dismutase as an
antioxidant in physiological stress. Physiologia Plantarun 76:456-459.
Penning de Vries F W T (1975) The cost of maintenance processes in plant cells. Annals Rot. 39: 77-92.
Pradet A, Bomsel J L (1978) Energy metabolism in plants under hypoxia and anoxia. Pages 89-118. In: Plant
life in anaerobic environments. D.D.Hook and R.M.M. Crawford, eds. Ann Arbor Science Publisher Inc,
Michigan.
Rea P A, Poole R J (1993) Vacuolar H+ translocating pyrophosphatase. Annu Rev Plant Physiol. Plant
Mol.Biol. 44: 157-180.
Raymond P, Pradet A (1980) Stabilization of adenine nucleotide ratios at various values by an oxygen limitation
of respiration in germinating lettuce (Lactuca sativa) seeds. Biochem. J. 190:39-44.
Robards A W, Lucas W J (1990) Plasmodesmata. Annu. Rev. of Plant Physiol. Plant Mol. Biol. 41: 369-419.

29

Setter T L, Kupkanchanakul T, Kupkanchanakul K, Bhekasut P, Wiengweera A, Greenway H (1987)

Concentration of CO 2 and O2 in floodwater and in internodal lacunae of floating rice growing at 1-2metre
water depths. Plant Cell Environ. 10:767-776.
Setter T L, Kupkanchanakul T, Waters I, Greenway H (1988a) Evaluation of factors contributing to diurnal
changes in floodwater in deep water rice fields. New Phytologist 110:151-162
Setter T L, Kupkanchanakul T, Kupkanchanakul K, Rhekasut P, Wiengweera A.Greenway H (1988b).
Environmental factors in deepwater rice areas in Thailand: O2, CO2 and ethylene. Pages 55-66. In:
Symposium of the 1987 International Deepwater Rice Workshop. IRRI, Los Banos., Philippines.
Setter T L, Waters I, Wallace I, Bhekasut P, Greenway H(1989) Submergence of rice I. Growth and
photosynthetic response to CO2 enrichment. Aust.J.Plant Physiol. 16:251-263.
Setter T L, Ingram K T, Tuong T P (1995) Environmental characterisation requirements for strategic research
in rice grown under adverse conditions of drought, flooding or salinity. Pages 3-18. In : Rainfed lowland
rice-agricultural research for high risk environments. KT Ingram, ed. International Rice Research Institute,
Los Banos, Philippines.
Setter T L, Ella E S, Valdez A P (1994) Relationship between coleoptile elongation and alcoholic fermentation
in rice exposed to anoxia. II.Cultivar differences. Ann.Rot. 74:273-279.
Setter T L, Ramkrishnaya G, Maurya Ram P C, Singh B R (1994) Environmental characteristics of floodwater
in flood-prone areas of northern and eastern India: Relevance to flooding tolerance of rice. Indian J.Plant
Physiol. 38: 34-40.
Stitt M (1990) Fructose-2-6-bisphosphate as a regulatory molecule in plants. Annu Rev. Plant Physiol.Plant
Mol.Biol. 41:153-185.
Thomson C J, Greenway H (1991) Metabolic evidence for stelar anoxia in maize roots exposed to low O2
concentrations. Plant Physiology 96:1294-1301.
Van Raalte M H (1940) On the oxygen supply of rice roots. Ann. Bot. Gardens Buitettzotg 50:99-113.
Van Toai T T, Bolles C S (1991) Postanoxic injury in soybean (Glycine max)seedlings.Plant Physiology
97:588-592
Waters I, Armstrong W, Thomson C J, Setter T L, Adkins S, Gibbs I, Greenway H (1989) Diurnal changes
in radial oxygen loss and ethanol metabolism in roots of submerged and non-submerged rice
seedlings. New Phytologist 113:439-451.
Webb T, Armstrong W (1983) The effects of anoxia and carbohydrates on the growth and viability of rice,
pea and pumpkin. J.Exp.Bot. 34:579-603.
Yocum, C.S., Hacket D.P., (1957) Participation of cytochromes in the respiration of the aroid spadex. Plant
Physiology 32:186-191.
Zhang F, Lin J J, Fox T C, Mujer C V, Rumpho M E, Kennedy R A (1994) Effect of aerobic priming on the
response of Echinochloa crus-pavonis to anaerobic stress: Protein synthesis and phosphorylation. (Submitted
to Plant Physiology).
Zhang Q, Greenway H (1993) Energy dependent transport in anoxic plant tissues. Pages 189-191. In:
Proceedings of the Plant Nutritions Colloqium. N.J. Harrow ed. Kluwer Academic Publishers, Perth,
Australia.
Zhang O, Lauchli A, Greenway H (1992) Effects of anoxia on solute leakage from beetroot storage tissue.
J.E.xp.Bot 43: 897- 905.
Zhang Q, Greenway H (1994) Anaerobic catabolism of aged beetroot storage tissue. J. Exp. Bot. (in press).

30

TOLERANCE TO SUBMERGENCE AND

ANAEROBIOSIS : RICE SEEDLINGS HAVE SOMETHING
STILL TO LEARN
Michael B. Jackson1 , Deborah M.E. Pearce 2 and
Jacky E. Summers 1
SUMMARY
Rice seeds germinate without oxygen, and the emerging coleoptile may
elongate to a greater lengths without oxygen than in its presence. In presence of
oxygen the coleoptile responds positively in enhancing carbon dioxide and ethylene
production which is expected under these conditions. However, germinating rice
fails to produce a root anaerobically, the coleoptile is limited in length to a few
millimeters and the first true leaves show low tolerance to oxygen shortage. When
totally submerged in clear water, small seedlings fail to gain dry mass through
photosynthesis and are dependent on seed reservers for survival. Other aquatic or
amphibious species, including Trapa natans (Water chestnut), Echinochloa oryzoides
(barnyard grass), and Potamogeton pectinatus (pond weed) appear to have overcome these
shortcomings.

INTRODUCTION
Higher plants must be well aerated in order to survive, compete and complete
their life cycle successfully. All plant parts require oxygen for respiration and other
biochemical oxidations/ oxygenations. Besides, green shoots require an external
input of carbon dioxide for photosynthesis that benefits the whole plant. Plants also
need to dispose the potentially active or damaging volatile metabolites such as
photosynthetic oxygen, acetaldehyde, ethylene or methyl jasmonate produced by
various metabolic pathways. The principal environmental factor affecting aeration
is water. Little available water adversly affects the growth whereas, too much water
can asphyxiate and may kill growing seedlings in majority of cases. Slow diffusion
of gas in water compared to air (D water / Dair) is approximately 1.13 10-4 which is
responsible for this traumatic effect of an otherwise harmless substance.
Despite these caveats, a sizable number of species tolerate flooded and
submerged conditions well to a large extent and constitute a vigorous and
ecologically significant aquatic or amphibious flora in many parts of the World
(Maltby, 1991). Rice ( Oryza sativa L.) is the only crop plant that can easily be placed
in this group and hence its dietary preeminence in the humid tropics.

1 Department of Agricultural Sciences, University of Bristol, Long Ashton Research Station, Bristol BS18 9AF,
Great Britain
2 School of Biological Sciences, Queen Mary and Westfield College, University of London, London E14NS,
Great Britain.

The physiological and metabolic characteristics of flood-tolerant plants are

varied and differ depending upon the stage of development and the species.
Tolerance often is achieved through a combination of short-term resistance to absence
of oxygen, physiological responses such as fast elongation under water that minimise
the duration of submergence or anoxia, and the ability to fix carbon dioxide
photosynthetically or bicarbonate dissolved in flood water. Tolerant plants must also
be able to survive the potentially damaging effects of the entry of oxygen into
previously anaerobic conditions (i.e., post anoxic injury - Andreev et al., 1991.) The
extent to which rice possesses these characteristics is reviewed in the present paper.

ANAEROBIC GERMINATION
Rice is well-known for its ability to germinate without oxygen. In this regard
it has few rivals only since the number of other species known to germinate
anaerobically is small. These include Erithrina caffra, Nuphar luteum and Scripus
mucronatus (Menegus et al., 1992a), and various species of Echinochloa some of which
are weeds in rice field (Barrett and Seaman, 1980). In each case only the shoot emerges
from the anoxic seeds. Thus, anaerobically germinating seedling is without an
anchoring rootlet. This can be a problem in fields to direct-sown rice where anaerobic
conditions make it difficult for the crop to become firmly established. The reason
for the failure of the emerging embryonic root under anaerobic condition is not clear,
however it survives, and quickly elongates when oxygen is made available (Kordan,
1977). However, one species, Trapa natans (the water chestnut), is able to germinate
by radicle emergence (Menegus et al., 1992a). This is the only species known where
root growth is not inhibited by anoxia. On the contray in seedling, root extension is
actually promoted by the absence of oxygen in this plant. The metabolic and
physiological basis of this unique characteristic are not yet fully understood. Metabolic
studies (Menegus et al., 1992a) of T. natans have shown a high rate of alcoholic
fermentation (20 u mol ethanol g-l FW h-1 ), the availability of respirable starch reserves,
and an ability to avoid cytoplasmic acidification achieved through. (i) minimal
production of potentially fatal lactic acid, (ii) proton consumption in conversion of
malate to succinate, (iii) arginine to putrescine, (iv) glutamate to Y-aminobutyric acid
(GABA). These metabolic features are strongly reminiscent of the rice coleoptile
(Menegus et al., 1992b; Fan et al., 1993; Perata et al., 1992; Ricard et al., 1991). They are
considered to help in avoiding acidification of the cytoplasm, maintain sufficiently high
levels of ATP and dissolved solutes to support cell expansion and membrane integrity,
while other energy consuming processes such as the overall levels of protein synthesis
(Mocquot et al., 1977) and fatty acid production (Vartapetian et al., 1978) are suppressed.
However, a large number of proteins can be synthesised anaerobically in the coleoptile
(Chirkova and Hoang 1981; Atwell and ApRees, 1986), and two dimensional gels of
radio- labelled proteins synthesised in aerobic and anaerobic conditions show
remarkably similar patterns (Ricard and Pradet, 1989).
One way of improving the performance of germinating rice seeds would be
to induce the embryonic root axis to elongate as a coleoptile, while retaining the
positive gravitropism typical to aerobic rice roots. This may be possible because
anoxia and root growth are not mutually incompatible in Trapa natans. A large
32

screening programme might be useful to have this capability in rice. Genetic

engineering is an alternative if differences in gene expression can be shown to
understand the contrasting behaviour of root and shoot. The number of proteins
synthesised by anaerobic roots is small compared to the coleoptile (Ricard and
Pradet, 1989) but the nature of anaerobic energy metabolism in roots and shoots of
rice seedlings appears remarkably similar (Mayne and Kende, 1986), at least in the
short term when switched from aerobic to anaerobic conditions (Fig. 1). This suggests
that deficiencies in fermentative sugar metabolism do not explain why roots fail to
emerge from the seed anaerobically, and that some other phenomenon in blocking
growth is involved. The absence of cell division in anoxic roots (Amoore 1961;
Kordan, 1976) may be one although it is notable that anoxia does not entirely stop
cells from dividing in the emerging rice coleoptile (Kordan 1976; Opik, 1973) or in
stems of P. pectinatus (Summers and Jackson, 1993). However, studies of the so-called
'gamma plantlets', in which cell divisions in the mature embryo of wheat Triticum
(
vulgare) or maize ( Zea mays ) are prevented by prior irradiation, show that, on
imbibition, the primary rootlet can still extend 9- 15 times its original length without
the need for cell divisions (Foard and Haber, 1961). But, anoxic rice seeds fail to
achieve even this limited amount of root extension, indicating that some other
process inhibits the radicle emrgence. Identifying the physiological and genetic basis
of this blockage would be an important step in achieving the goal of producing rice
seeds in which anaerobic germination incorporates root emergence.

Fig. 1.

Comparison of the production of radioactive carbon dioxide by root or coleoptile of rice

supplied with [U- 14 C] glucose at time zero. After 3 h (arrow) the atmosphere was changed
from aerobic to anaerobic. (After Mayne and Kende, 1986).

33

GROWTH OF COLEOPTILE WITHOUT OXYGEN

Anaerobic germination can lead to successful development of the seedling
if some source of oxygen is made available before reserves in endosperm run out.
This is usually the surface water which is well aerated or the atmosphere where
the coleoptile allows oxygen to penetrate and move the entire seedling (the
Schnorkel effect - Kordan, 1974) through the hollow coleoptile (Kutschera et
al., 1990) and highly permeable aerenchymatous leaf and root interiors (Fig. 2).
Aerenchyma is created constitutively by the collapse of cells and in roots, it can
be initiated within a day of their leaving meristem and involves cell wall
degradation (Webb and Jackson, 1986). Additional aerenchyma may be induced
by the action of ethylene in some cultivars (Justin and Armstrong, 1991), not in
all (Jackson et al., 1985).
Fast elongating coleoptile has greater chances to reach the oxygenated
zones closer to the surface. Thus the rice coleoptile not only survives and even
grows in length without oxygen for serveral days, but the rate of extension can
also be stimulated in the absence of oxygen. In Fig.3 it has been shown that in

Fig. 2.

34

Extension growth by the coleoptile of rice and Echinochloa oryzoides seedlings grown in air
(Air), air containing 1.0 Pa ethylene (Ethylene), or in nitrogen gas (No oxygen) for up to
8 d (taken from Pearce and Jackson, 1991).

the submergence tolerant rice cultivar 'Calrose' the length of coleoptile was 3.4 cm
longer in nitrogen environment than air after 8 day. This growth is unlikely due to
endogenous hormones such as a uxin, ethylene or gibberellins since they are inactive
when supplied to anoxic coleoptiles (Jackson and Pearce 1991; Horton, 1991).
However, the diamine putrescine appears to play a positive role in anaerobic
coleoptile elongation (Reggiani et al., 1989) while abscisic acid could be inhibitory
(Horton, 1991). Futhermore, interference in the already weak gravitropism in
anoxic
rice
coleoptile
by
auxin
transport
inhibitor
such
as
N-1
naphthylphthalamic acid (recently reported by Horton, 1994) implies that
endogenous auxin has some role under oxygen-free conditions. However, it is
contrary to competing weed E. oryzoides, where coleoptile elongation is neither
stimulated nor inhibited in absence of oxygen (Fig. 3) In these species, a 20- mm
long coleptile can easily be obtained regardless of the availability of oxygen or
the presence of elevated levels of ethylene (Fig. 3) or carbon dioxide (Pearce and

Fig. 3.

Anaerobic growth by shoots of Potamogeton pectinatus tubers. A. Stem extension of shoots

submerged in water sparged with air ('Air') or with an anaerobic gas mixture of hydrogen
and oxygen ('No oxygen'). B. Negative (i.e, upward) gravitropic curvature by shoots in
submerged or in a gas phase with ('Aerobic') or without (Anaerobic') oxygen. (Summers and
Jackson, 1993).

35

Jackson, 1991). Thus, such indifference to extremes in the gaseous environment is

unprecedented.
The positive effect of the absence of oxygen on elongation in rice coleoptiles
is small compared to those reported for the stem of the overwintering tubers of a
temperate aquatic weed Potamogeton pectinatus. This species is the only known
example where shoot elongation is stimulated by the complete lack of oxygen
(Summers and Jackson, 1993). In this case it is the stem tissue that reponds, but
with a vigour which far exceeds the rice coleoptile (Fig. 4). Further, growth may
proceed for 14 d and is highly responsive to gravity (Fig. 4), suggesting that
elongation is oriented upwards and towards the better aerated surface zones. In
contrast, the anaerobic rice coleoptile has an impaired ability to respond to gravity
(Kutchera et al., 1990; Horton, 1994) resulting in disorientated extension that may
affect the adaptive value of the stimulation caused by anoxia to extension growth
(Kordan, 1975).
COLEOPTILE EXTENSION IN PRESENCE OF DISSOLVED OXYGEN
When sumerged, coleptiles of rice move towards water that contains some
dissolved oxygen, several additional mechanisms are involved for the continuous and
unabated coleoptile elongation. The presence of some oxygen permits ethylene
biosynthesis which would have rather been blocked during the preceeding anaerobic
condition because the oxidation of ethylene precursor 1- aminocyclopropaneI-carboxylic acid (ACC) requires free oxygen. When oxygen is admitted, initially
ethylene production is unusually fast because the accumulated ACC is converted to
ethylene (Table 1).
Table 1.

Effect of transferring seedling of rice (Oryza sativa) from anaerobic to

aerobic conditions on coleoptile length, ethylene production and
concentrations of 1-aminocyclopane-1- carboxylic acid .(ACC). From
Pearce et al., (1992).

Condition

Coleoptile
length
(mm)

Ethylene
production
(nl g-1 h-1)

ACC
(nmol g-1
fresh wt.)

No oxygen for 4 d

29.7

00

23.1

Air for 4 d

18.7

2.1

02.5

No oxygen " air on

day 3

32.0

3.7

11.7

The surrounding water then entraps the newly synthesised ethylene (Ishizawa
and Esashi, 1984), and the presence of some oxygen also allows cells to respond
36

physiologically to the ethylene. In case of coleoptile cells, elongation is enhanced

(Table 1 and Pearce and Jackson, 1991) by ethylene in association with the action of
auxin (Ishizawa and Esashi, 1983). The rice coleoptile was the first organ, of any
specices, found to elongate faster in respone to ethylene gas (Ku et al., 1970). This
positive response of gas (optimum 0.1 Pa) can be related to the presence of ethylene
binding sites (Sander et al., 1990) and to enhanced translocation of sucrose into the
coleoptile (Ishizawa and Esashi, 1988). Ethylene action is augmented by other features
of the submerging environment also. These include elongation promoting influences of
carbon dioxide (optimum 10-15 kPa), partial oxygen shortage (optimum 3 kPa) and the
stretching influence of buoyant tension (Pearce and Jackson, 1991). Ethylene, carbon
dioxide and low oxygen act either separately (Raskin and Kende, 1983), or interactively
(Ku et al., 1970). However, neither low oxygen nor high carbon dioxide act by stimulating
ethylene biosynthesis (Raskin and Kende, 1983; Pearce and Jackson, 1991) and
presumably possess intrinsic growth promoting activity in their own right. In addition
to promoting elongation, the extra post-anoxic ethylene conceivably could help in
reducing injury (Ushimaru et al., 1992) arising from hydrogen peroxide generation
(Mehlhorn 1990) by promoting the activity of ascorbate peroxidase.
Emergence of leaves and stems
When the elongation of coleoptile stops, the first true leaves, enclosed by the
coleoptile, emerge in presence of some amount of oxygen. Leaf elongation is strongly
inhibited in abscence of oxygen in rice (Turner et al., 1981). Thus, further growth
underwater is not possible unless a source of oxygen is met. This is probably true
for majority of other species. However there can be exceptions. For example, leaves
on shoots of P. pectinatus tubers extend vigorously in absence of oxygen, although
the rate is slower than in well-aerated water (Table 2).
Table 2.

Effect of anaerobic conditions on elongation by the first leaf of

overwintering tubers of Potamogeton pectinatus, over 5 d. From Summers
and Jackson, unpublished.
Condition

Water

Gas phase

Aerobic

6.2*

4.5

Anaerobic

3.3

3.9

Significantly different from anaerobic (p= 0.05)

If oxygen is present alongwith light, extension in one of the young rice

is promoted by total submergence (Fig.5), while senescence in the first leaf
seedings is promoted. Both effects are probably mediated by ethylene
submerged shoots become enriched with endogenous ethylene (Jackson et al.,

leaves
of the
since
1987)
37

whereas application of ethylene to non-submerged plants stimulate extension in the

youngest visible leaf (Fig. 5; Ku et al., 1970; Jackyon et al., 1987, but see Raskin and
Kende, 1984) and promote loss of chlorophyll in the oldest leaf (Jackson et al., 1987).
Unexpectedly, the effect of ethylene is linked with increased dry weight (Jackson et
al., 1987 and Fig. 5). This probably indicates the ability of ethylene to increase the
assimilate import from the seed (Ishizawa and Esashi, 1988). The promoting effects
of submergence, and ethylene on leaf extension are not shared by the competitive
weed i.e. E. oryzoides (Pearce, 1990) but can be found in some other species e.g., P.
pectinatus (Summers and Jackson, 1993). Elongation in the roots of rice is also
favoured by modest increases in ethylene (Fig.5, Smith and Robertson, 1971; Konings
and Jackson, 1979). As noted in coleoptile, carbon dioxide entrapped in submerged
condition also enhances the leaf extension in rice (Raskin and Kende, 1984). In the
dark this may suggest growth regulating activity of carbon dioxide, but in light it
is due to increased photosynthetic carbon fixation rate (Setter et al., 1989).

Fig. 4.

38

Effect of complete submergence or of ethylene treatment given for 3 d to 7-d-old light-grown

seedlings of rice. Effect of submergence (A) or 10 Pa ethylene (B) on elongation by the
longest root and the second and third oldest leaves. Effect of ethylene on growth in dry
weight (C). From Pearce, 1990.

Increase in dry mass

A close examination of the dry mass gain in 7-d old rice plants submerged in
the light (850 u mol m-2 s-1) shows that the shoot and root can gain only a small
amount of dry mass over 3d. This is achieved at the expense of the seed reserve
(Table 3), indicating just survival with little or no net gain from external carbon. The
decrease in photosynthetic rate is not due to damage of the photosynthetic machinery
since submerged leaves release oxygen-rich gas bubbles in light and the surrounding
water becomes enriched with dissolved oxygen. Others have shown that submerged
rice plants, if supplied with carbon dioxide, fix it photosynthetically showing high
growth rates (Setter et al., 1989). Thus, it is clear that submerged rice seedlings can
and do fix carbon photosynthetically, but mostly utilize the respiratory released
carbon dioxide. However, E. oryzoides is a contrasting type as its seeds are very small
and have small amount of reserves (Table 3).
Table 3.

Effect* of 3 days complete submergence and 3 days recovery

(de-submergence) on shoot, root and seed dry weights of rice (cv.
Calrose) and Echinochloa oryzoides seedlings.

Day

Shoot
wt(mg)

Rice
Root
wt(mg)

Seed
wt(mg)

Shoot
wt(mg)

7.4

3.2

11.8

5.0

3.7

2.4

Not Submerged

14.4

5.3

6.6

22.0

9.6

2.1

Submerged

8.9

2.6

5.6

6.9

3.0

2.3

24.7(15.1) 2.1(0)

Day 0

Echinochloa
Root
Seed
wt(mg)
wt(mg)

Day 3

Day 6
Not submerged

20.1(5.7) 13.4(8.1) 2.9(-3.7)

33.0(11)

De-submerged

13.4(4.5) 6.5(3.9)

30.5(23.6) 21.9(18.9) 1.6(-0.7)

6.2(0.6)

Plants were 7-d old at the start; Figures in parentheses are changes over the last 3 days of the
experiment.

Besides, seedlings submerged in light survive and gain a small increase in dry
weight. E. oryzoides has a superior ability to fix carbon dioxide (Smith and Walker,
1980). This may be expected from a C4 species compared to a C3 plant rice because
a C4 plant has low compensation point (Madsen, 1993). The difference between the
two species in their dependence on external carbon dioxide is shown in Fig 5. If
seedlings of both the species are grown with their shoots in moist air from which
carbon dioxide has been removed completely, growth in dry mass by the root and
39

Fig. 5.

Effect of absorbing carbon dioxide from air on growth in dry mass by initially 7-d old light-grown
seedling with the roots in nutrient solution and the shoots in a gas phase flowing at approximately
2L min -1 at 850 u mol m -2 s-1 PAR. Figures above the bars are relative growth rates (mg
mg-1 d-1 ). From Pearce, 1990.

shoot stops in E. oryzoides. But in rice, root and shoot growth continues but at the
expense of seed reserves. Thus, submerged E. oryzoides achieves its small increase
in mass by fixing carbon dioxide from the water and not by using seed reserve. In
contrast, rice while being unable to fix carbon dioxide from the dissolved water,
compensates it by drawing from the seed reserves. This way both the species survive
three days total submergence in well-illuminated water at the 7-leaf stage and are
able to grow strongly if shoots are exposed to air (Table 3).

40

CONCLUSION
Rice can germinate without oxygen, and its seedlings can grow in submerged
conditions that are partially depleted of oxygen. The coleoptile elongates to a greater
length than in normal in response to anaerobiosis, partial oxygen shortage, and
enriched ethylene or carbon dioxide. These factors are inhibitory for most of the
species. However, roots fail to emerge from rice seeds unless some oxygen is present,
while anaerobic stimulation of coleoptile elongation is limited to only a few
millimetres. Forthermore, extension of the first true leaves is dependent on oxygen,
although its effect is increased by ethylene and carbon dioxiode . If the shoots of
small seedlings are submerged, photosynthetic fixation of carbon dioxide from the
flood water is minimal and the shoot and roots depend on seed reserves for survival
over at least for 3 days.
Certain other species are shown to perform better than rice in several aspects.
Trapa natans can produce a root when germinating anaerobically and stems of
anaerobic Potamogeton pectinatus tubers elongate much more vigorously for longer
period than rice coleoptiles and have stronger negative gravitropism. The coleoptile
of Echinochloa oryzoides can extend its full length irrespective of presence or absence
of oxygen whereas the leaves of P. pectinatus can elongate strongly with oxygen. The
shoots of E. oryzoides are able to survive independently of the seed reserves when
submerged for 3 days and grow more vigorously than rice after exposed to resubmergence. Studies show how these species out-perform rice under conditions of
poor aeration, and ultimately, may help select or create strains of rice with superior
performance in flooded conditions. These studies show the capabilities of the species
other than rice to perform better under the poor condition of aeration and thus open
the scope for developing better strains of rice with such characteristics to perform
well under flooded conditions.
REFERENCES
Amoore JE (1961) Arrest of Mitosis in roots by oxygen lack or cyanide. Proc. Roy. Soc. B. 154 :95-108.
Andreev VY, Generozova I P, Vartapetian B B (1991) Energy status and mitochondrial
ultrastructure of excised pea root at anoxia and post anoxia. Plant Physiol. Biochem. 29:171-176.
Atwell B J, ApRees T (1986) Distribution of protein synthesised by seedlings of Oryza sativa grown in
anoxia. J. Plant Physiol. 123: 401-408.
Barrett S C H, Seaman D E (1980) The weed flora of California rice fields. Aquatic Bot. 9: 351-376.
Chirkova T V, Hoang K L (1981) Electrophoretic pattern of cytoplasmic and mitochondrial proteins in
wheat and rice roots under anaerobiosis. Fiziol. Rast. 28:136-143,
Fan T W M, Lane AN, Higashi R M (1993) Energy and fermentation metabolism in hypoxic rice coleoptile
- a multinuclear NMR approach, Pages 333-352 In: Interacting stresses on plant in a changing climate.
M B Jackson and C R Black, eds. Springer- Verlag, Berlin, Germany.
Foard D E, Haber A H (1961) Anatomic studies of gamma-irradiated wheat growing without cell division.
Amre. J. Bot. 48:438- 446.
Horton R F (1991) The effect of ethylene and other regulators on coleoptile growth of rice under anoxia,
Plant Sci . 79:57- 62.
Horton R F (1994) Gravitropism in rice seedlings: the effects of anoxia and N-1 naphthylphthalamic acid.
Plant Sci. 95:41- 47.

41

Ishizawa K, Esashi Y (1983) Co-operation of ethylene and auxin in the growth regulation of rice coleoptile
segments. J.Exp.Bot. 34: 74-82.
Ishizawa K, Esashi Y (1984) Gaseous factors involved in the enhanced elongation of rice coleoptiles under
water. Plant Cell Environ. 7 : 239-25.
Ishizawa K, Esashi Y (1988) Action mechanism of ethylene in the control of sugar translocation in
relation to rice coleoptile growth I. Sucrose metabolism. Plant Cell Physiol. 29 : 131-141.
Jackson M B, Pearce D M E (1991) Hormones and morphological adaptation to aeration stress in rice.
Pages 47-67. In: Plant life under oxygen deprivation. Ecology, physiology and biochemistry. M.B. Jackson,
D.D. Davies and H. Lambers, eds. SPB Academic Publishing, The Hague, The Netherlands.
Jackson M B, Fenning T M, Jenkins W (1985) Aerenchyma (gas space) formation in adventitious roots of
rice ( Oryza sativa L.) is not controlled by ethylene or small partial pressures of oxygen. J. Exp. Bot. 36:
1566-1572.
Jackson M B, Waters I, Setter T, Greenway H (1987). Injury to rice plants caused by complete submergence;
a contribution of ethylene J. Exp. Bot. 37 : 1826-1838.
Justin S H F, Armstrong W (1991) Evidence for the involvement of ethylene in aerenchyma formation in
adventitious roots of rice ( Oryza sativa L.). New Phytol. 118: 49-62.
Konings H, Jackson M B (1979) A relationship between rates of ethylene production by roots and the
promoting or inhibiting effects of ethylene on root elongation. Z Pflanzenphysiol. 92: 385-397.
Kordan H A (1974) Patterns of shoot and root growth in rice seedlings germinating in oxygen. J.Plant
Ecol. 11: 685- 690.
Kordan HA (1975) Relationship between oxygen availability and transverse and vertical shoot geotropism
during germination of submerged rice seedlings. Ann. Bot. 39: 249-256.
Kordan H A (1976) Mitotic activity in rice seedling under oxygen deficiency. J. Cell Sci. 20 57-59.
Kordan H A (1977) Latent effect of anaerobiosis on root growth in germinating rice seedlings. Plant Sci.
Lett. 9 53-56.
Kordan H A, Ashraf M (1990) Environmental anoxia is unnecessary for inhibiting chloroplast
photomorphogenesis in rice coleoptiles ( Oryza sativa L.). J. Exp. Bot. 41 : 435-440.
Ku H S, Suge H, Rappaport L, Pratt H K (1970) Stimulation of rice coleoptile growth by ethylene. Planta
90 : 333-339.
Kutschera U, Siebert C, MasudaY, Sievers A (1990) Effects of submergence on development and
gravitropism in the coleoptile of Oryza sativa L. Planta 183: 112-119.
Madsen T V (1993) Inorganic carbon assimilation and growth of aquatic macrophytes. Pages 267-285. In:
Interacting stresses on plants in a changing climate. M.D. Jackson and C.R. Black, eds. Springer-Verlag,
Berlin, Germany
Maltby E (1991) Wetlands-their status and role in the biosphere. Pages 3-21 In: Plant life under oxygen
deprivation. Ecology, physiology and biochemistry M.R. Jackson, D.D. Davies and H. Lambers, eds. SPB
Academic Publishing, The Hague, The Netherlands.
Mapelli S, Bertani A (1993) Endogenous phytohormones and germination of rice under anoxia: indoleacetic
acid and abscisic acid. Pages 353-363. In: Interacting stresses on plants in a changing climate. M.B. Jackson,
C.R. Black, eds. Springer-Verlag, Berlin, Germany
Mayne G R, Kende H (1986) Glucose metabolism in anaerobic rice seedlings. Plant Sci. 45:31-36.
Mehlhorn H (1990) Ethylene-promoted ascorbate peroxidase activity protects plants against hydrogen
peroxide, ozone and paraquat. Plant Cell Environ. 13:971-976.
Menegus F, Cattaruzza L, Scaglioni, Ragg E (1992a) Effects of oxygen level on metabolism and development
of seedlings of Trapa natans and two ecologically related species. Physiol. Plant. 86 : 168-172.
Menegus F, Cattaruzu L, Molinari H, Ragg E (1992b) Rice and wheat seedlings as plant models of high
and low tolerance to anoxia. Pages 53-66. In: Surviving hypoxia: Mechanisms of control and adaptation.
P.W. Hochachka, ed. CRC Press, Boca Raton, USA.

42

Mocquot B, Pradet A, Litvak S (1977) DNA synthesis and anoxia in rice coleoptiles. Plant Sci. Lett. 9 :
365-371.
Opik H (1973) Effect of anaerobiosis on respiratory rate, cytochrome oxidase activity and mitochondrial
structures in coleoptiles of rice (Oryza sativa. L.). J.Cell Sci. 12 : 725-739.
Pearce D M E (1990) A comparison of the responses of seedlings of rice and Echinochloa oryzoides to
submergence. PhD dissertation, University of Bristol, Great Britain.
Pearce D M E, Jackson M B (1991) Comparison of growth responses of barnyard grass ( Echinochloa oryzoides)
and rice (Oryza sativa ) to submergence, carbon dioxide and oxygen shortage. Ann. Bot. 68:201-209.
Pearce D ME, Hall KC, Jackson M B (1992) The effect of oxygen, carbon dioxide and ethylene on ethylene
biosynthesis in relation to shoot extension in seedlings of rice ( Oryza sativa) and barnyard grass ( Echinochloa
oryzoides ). Ann. Bot. 69: 441- 447.
Perata P, Pozueta J, Akazawa T, Yamaguchi J (1992) Effect of anoxia on starch breakdown in rice and
wheat seeds. Planta 188: 611-618.
Perata P, Alpi A (1993) Plant responses to anaerobiosis. Plant Sci. 93:1-17.
Raskin I, Kende H (1983) Regulation of growth in rice seedlings. Pl. Growth Regul 2: 192-203.
Raskin I and Kende H (1984) Regulation of growth in stem sections of deepwater rice. Planta 160 :66-72.
Reggiani R, Hochkoeppler A, Bertani A (1989) Polyamines and anaerobic elongation of rice coleoptile.
Plant Cell Physiol. 30:893-898.
Ricard B, Pradet, A (1989) Anaerobic protein synthesis in different organs of germinating rice seedlings.
Plant Physiol, Biochem. 27: 761-768.
Ricard B, Rivoal J, Spiteri A (1991) Anaerobic stress induces transcription and translation of sucrose
synthase in rice. Plant Physiology. 95: 669-672.
Sanders I O, Ishizawa K, Smith A R, Hall M A (1990) Ethylene binding and action in rice seedlings. Plant
Cell Physiol. 31 :1091-1099.
Setter T L, Waters I, Wallace I, Bhekasut P D, Greenway H (1989) Submergence of rice I. Growth and
photosynthetic response to CO2 enrichment of floodwater. Aust, J. Plant Physiol. 16: 251-263.
Smith K A, Walker N A (1980) Photosynthesis by aquatic plants: effects of unstirred layers in relation to
assimilation of CO2 and HCO -3 and to carbon isotope discrimination. New Phytol. 86: 245-259.
Smith K A, Robertson P D (1971) Effect of ethylene on root extension of cereals. Nature 234: 148-149.
Summers J, Jackson M B (1993) Promotion of stem extension in an aquatic monocot (Potamogeton pectinatus
L.) by the complete absence of oxygen, and by partial oxygen shortage. Page 315-326. In: M B Jackson
and C R Black, eds. Springer-Verlag, Berlin, Germany.
Turner F T, Chen C-C, McCauley G N (1981) Morphological development of rice seedling in water at
controlled oxygen levels. Agron. J. 73 : 566-570.
Ushimaru T, Shibaska M, Tsuji H (1992) Development of the O2 detoxification system during adaptation
to air of submerged rice seedlings. Plant Cell Physiol. 33 :1065-1071.
Vartapetian B B, Mazliak P, Lance C (1978) Lipid biosynthesis in rice coleoptiles grown in the presence
or in the absence of oxygen. Plant Sci. Letts. 13 : 321-3213.
Webb JD, Jackson M B (1986) Transmission and cryo-scanning electron microsocopy study of the formation
of aerenchyma (cortical gas-filled space)in adventitious roots of rice) Oryza sativa) J Exp. Bot. 34: 579-603.

43

ENVIRONMENTAL AND PLANT MEASUREMENT

REQUIREMENTS FOR THE ASSESSMENT OF DROUGHT,
FLOOD AND SALINITY TOLERANCE IN RICE
P.C. Ram 1 , B.B. Singh 1 , A.K. Singh 1, V.K. Singh1
O.N. Singh 1 , T.L. Setter 2 , R.K. Singh 1 and V.P. Singh2
SUMMARY
For the assessment of stress tolerance of rice, a set of environmental and plant
measurements are presented which need to be monitored prior to and during the
different growth and developmental phases of plants. These measurements on soil,
plant, atmosphere and flood water systems are essential for integration and
characterization of rice ecosystems which may facilitate interpretation of stress
tolerance mechanism of plants.
The findings of experiments done at the Faizabad consortium site alongwith
a brief review are presented in order to prioritize and establish relationships between
environmental measurements and rice productivity under adverse ecosystems of
rainfed lowland rice culture.
INTRODUCTION
Diversity and variability of environment in different rice growing countries
and regions (Herdt and Barker, 1977) would make the development of an appropriate
technology difficult. Herdt and Barker (1977) have listed a set of environmental
factors that vary across sites, seasons and years. With the available literature on soils
and climate, information is still lacking on many rice-growing environments.
The average rice yields in different countries range from less than 1 t/ha to
more than 6 t/ha. There are a number of biological, environmental and
socio-economic reasons for the large differences in rice yields associated with rainfed
lowland, deepwater and upland rice in tropical and temperate regions. Of these,
temperature, solar radiation and rainfall influence rice yield most by directly
affecting the physiological processes involved in grain production. Yield is also
indirectly affected through diseases and insects. These factors are often difficult to
separate from one another in the field and influence crop period, productivity and
stability which are the important aspects of rice cultivation. The complete
information on environmental measurements are seldom collected along with plant
measurements, hence the interpretation of data becomes difficult; and at times,
inaccurate and misleading conclusions are drawn. Further, such experimental data
can not be extrapolated to other sites or times unless essential environmental and
1
2

N.D. University of Agriculture and Technology, Kumarganj, Faizabad,U.P.,

International Rice Research Institute, P.O. Box 933, Manila, Philippines

India

plant measurements are generated and quantified accurately. Screening of rice

genotypes for flood and salinity tolerance should be done under well defined and
characterized sites. For example, saline areas may also have the alkalinity problems
along with high levels of soluble salts in the medium. Similarly for flooding, the
depth and duration of submergence, the growth stage of plant at the time of
submergence, turbulence and turbidity etc. may be important factors which affect
biomass and yield production. These environmental and plant measurements can
well be used for development of basic models to quantify growth and development
of rice under drought, flooding and salinity situations.
MATERIALS AND METHODS
The findings reported are from a number of experiments conducted under
drought, submergence or salinity conditions.
Drought
Three rice genotypes were tested during 1992 wet season for their relative
performance under drought imposed at vegetative and boot stages of plant growth.
The details of methodology are described below :
*

Culture : Pot experiment - Soil culture, 30 cm pot size

Genotypes - IR 28 (intolerant) IRAT -142 (medium tolerant) and Kachani

(tolerant)

Stages of water stress : Tillering (T) and Booting (B)

Method of stress : With-holding of irrigation until the appearance of

permanent wilting point (PWP).

Revival : After 48 hours at PWP the plants were revived by application

of water into the pots.

Observations
*

Soil water content : Gravimetrically at field capacity and PWP (Jackson,

1973).

Leaf water potential: Liquid equilibration method (Shardakov 1948).

Growth and Yield : Tillers, biological and grain yields, grain sterility etc.

Flooding/submergence
*

The performance of some of the elite rice lines was assessed under
prolonged partial submergence during 1992 and 1993 wet seasons under
field conditions.

No. of Genotypes - 201

Age of seedling at planting - 30 days

Water depth - Variable (monitored during the crop season)

45

Observations
*

Climatic parameters : Rainfall, sunshine hours, temperature,

humidity.

relative

Plant parameters : Survival, growth and yield parameters.

Water characterization of rainfed lowland rice fields: The measurement

of dissolved oxygen, temperature, CO2 and total inorganic carbon (TIC)
was done during 1993 wet season at crop research stations, Pagalabhari
and Ghagharaghat. Dissolved oxygen was measured using a submersible
polarographic O2 electrode (Syland model 610, Heppenheim, West
Germany) attached to a measuring rod. The electrode was calibrated in
the field using aerated water prepared by vigorously shaking distilled
water in air for 1-2 min. The data on temperature and dissolved O2
concentrations have been presented here while dissolved CO2 and total
inorganic carbon alongwith other parameters have been presented in
another paper (Setter et al., 1995).

Salinity/sodicity
The experiments were conducted in two situations: the normal soil which was
moderately alkaline (pH 8.5) and partially reclaimed sodic soil which had a pH 9.7.
Rainfed lowland rice genotypes were planted in field during 1992 (64 lines) and 1993
(55 lines). The details of experiments are given below :
*

No. of Genotypes : - 64 (1992), 55 (1993)

Seedling age at planting : 30-35 days

Number of seedlings planted : 3-4 seedlings/hill

Fertilizer applied : 120 , 60, 60 N P and K (Kg/ha)

Observations
*

Climatic parameters -Rainfall, sunshine, temperature (Max. Min.) relative

humidity during crop season.

Soil characteristics- Physical and chemical parameters (pH, ECe, organic

carbon, texture, ESP, CaCO 3 , N, P, K and Zinc) were monitored before
planting.

Plant parameters - A number of plant parameters were measured at

different growth stages in order to assess the performance of varieties and
possible mechanism of salt tolerance.

RESULTS AND DISCUSSION

Drought
Drought is a recurrent feature of rainfed rice cultivation limiting its growth
and productivity. Ironically, the timing, intensity, duration and distribution of rain
46

in many parts of rainfed rice growing areas arc uncertain, which cause mild to severe
water deficit during different phases of plant growth. About half of the 38 million
ha of the rainfed lowland rice ecosystem in the world has been classified either as
drought prone or drought and submergence prone (Garrity et al., 1986).
A number of soil parameters which affect plant establishment and growth
under drought prone rainfed lowland areas need to be measured. These include soil
matric potential, hydraulic conductivity, bulk density, field capacity, wilting point
etc.
Water availability and uptake by rice plants decrease as soil matric potential drops
below zero i.e. the matric potential of saturated soil. A careful monitoring of soil matric
potential and leaf water potential is required because ample reduction in leaf expansion,
tillering and leaf photosynthesis occurs long before leaf rolling and wilting, the two
common symptoms of drought stress. CO2 assimilation is Educed by soil water potential
of only -0.5 bar (OToole and Baldia, 7982). Generally, impact of water deficit was not
observed until the soil matric potential dropped below -1 bar and leaf water potential
below -1.5 bar when leaf rolling generally begins (Setter et al., 1993).
Table 1.

Essential and desirable data for environmental and plant characteristics

during drought, flash flooding and salinity.

(Abbreviations : S-single measurement; 2W-twice weekly; W- weekly; 2devery 2 days. d- daily; E-essential assuming equipments are available, D- desirable).
Measurements

2W

Frequency
W
2d

Stress
Drought Flooding

Salinity

1- SITE DESCRIPTION
a.

Altitude, Latitude and

Longitude

b. Topographical location
including slope

2- SOILS
a.

Soil profile description,

bulkdensity, F.C., wilting
point, saturation point and
soil texture

b. Soil chemical properties

including nutrients
c. Water table (Piezometer,
ground water tube)
d. Soil moisture or matric
potential within one day of
rainfall event, otherwise
weekly.

X
X

Continued ................

47

Measurements
S

2W

Frequency
W
2d

Drought

e. EC of soil (state depth

measured) and ground
water.

f.

pH of soil (state depth

measured)

Stress
Flooding
D

Salinity
E

g. Hydraulic conductivity/
percolation rate

h. Depth and thickness of

CaCO3 layer

3- FLOOD WATER
a. Flood water depth-rate
of increase or decrease

b. Turbidity of flood water

(monitor during event)

c.

Temperature of flood water

(monitor at mid afternoon
during an event)

d. Lateral movement of flood

water (monitor during event).

e.

pH (monitor during event)

f.

EC (including source of
flooding)

g.

O2 concentration (monitor at
water and soil surface and half
way within canopy; before
dawn and at mid afternoon)

h.

CO2 concentration(monitor
as per O2)

i.

Ethylene concentration
(monitor as per O2 )

4- CLIMATE (AT 0800H)

a.

Rainfall by experiment

E
E

b.

Air temperature (max. & min.)

c.

Radiation

d.

Wind speed (at 2m height)

e.

Air humidity or vapour

pressure

f.

Evaporation by site

Continued................

48

Measurements

2W

Frequency
W
2d

5-

PLANT CHARACTERISTICS AND PESTS

a.

Percent survival

b. Height, density and tiller

number
c. Biomass and grain yield
at maturity

d. Yield components

Stress
Drought Flooding

Salinity

e. Plant processes and

metabolic events
(phasic otherwise weekly)

f.

Leaf tip drying or rolling

or burning

g. Date of 50% panicle

emergence

X
X

h. Plant water balance - water

potential, osmotic potential
RWC, T P (weekly or phasic)

E
E

E
E

i.

Percent unfilled spikelets

j.

Root parameters (depth,

volume, and density) up
to flowering

k. Dry matter partitioning

(phasic otherwise single)

1. Disease, weeds and insects

damage (occurrence and
severity)

m. Siltation on the leaves

n. Leaf characters (Waxy, hairy,

glaborous etc.)

E
D

Exposure of plants to mild drought induced hardening of plants (Table 2,

Figures 1 and 2). Leaf roiling and tip drying were inversely related to leaf water
potential (r = -0.85 and -0.80 respectively; O Toole and Moya, 1978). Hence, these
factors are considered as essential plant measurements (Table 1). Soil types and their
physical conditions heavily influence growth and tillering of rice plants. Heavy
puddled soils exert strength and thus reduce tillering. If soil dries, cracks develop
even though soil water potential below 10-15 cm depth remains as high as -0.1 bar
(Castillo et al., 1992).

49

Plants can overcome drought effects either through avoidance or tolerance

(Levitt, 1980). Drought avoidance is the ability of plants to maintain high water
potential under reduced soil water availability and, infact, avoid dehydration.
Dehydration tolerance is the ability of plants to withstand internal water deficits
with minimal injury. In other words it is true protoplasmic resistance.
Another way to cope up with drought is escape. This is where the plant
completes its life cycle well ahead of the onset of drought. Crop duration is adjusted
so that critical stages like panicle emergence do not coincide with probable drought
periods. Early or late maturing, photoperiod-sensitive varieties may be able to escape
and recover quickly from drought. The exposure of rice genotypes to severe water
deficits at vegetative and boot stages depicted a contrasting response in the three
genotypes studied (Table 2). Leaf water potential, however, was initially similar in
all the lines at tillering, the appearance of PWP was at significantly lower water
potential in the tolerant genotype, Kachani (-21.2 bar) than in the susceptible
genotype, IR 28(-19.0 bar). On the removal of stress, Kachani revived quickly. Water
deficit imposed at boot stage either alone or in Combination with tillering, depicted
the same pattern for PWP and revival. Biological and grain yields and total grains
per panicle also reflected the adverse effects of water deficit (Figs. 1 and 2).

Fig. 1.

50

Effect of water deficits on yield in rice genotypes (per cent reduction over unstressed at top
of the bars.)

Fig. 2.

Total grains and per cent sterility (at the top of bars) in rice genotypes under water deficits.

Stress at tillering induced irrepairable loss in total biomass of all the genotypes,
but the grain yield was reduced more due to water deficit at booting stage. A
combined stress was more detrimental, but the effect was less severe in the tolerant
genotype, Kachani (Figures 1 and 2) and it was possibly due to appearance of PWP
at still lower water potential. Grain sterility was the main component responsible
for yield reduction which was 61% in IR 28 against only 32% in Kachani with stress
at tillering + boot stages. In order to quantify the stress response and explore the
mechanisms of tolerance/susceptibility to drought, the environmental parameters
listed in Table 1 need to be measured.
Effect of drought varies with the genotype, the phasic development at which
plants experience stress, the severity and the duration of drought. Rice yields are
most susceptible to water deficits at flowering than at vegetative phase (Ram et al.,
1988; Hsiao, 1982; Reyniers et al., 1982). Drought at vegetative stage causes
irrepairable loss of canopy whereas at flowering stage, water deficit hampers anthesis
and seed setting leading to higher spikelet sterility and lower yields (Ram et al.,
1988). Drought at flowering also reduced effective leaf area and photosynthesis, thus
plants have to depend on pre anthesis reserves which may impart tolerance against
internal water deficits (Chaturvedi and Ingram, 1988; Austin et al., 1980).
Furthermore, if stress coincides with anthesis, the panicle excertion may be inhibited,
spikelet sterility occurs (Ekanayake et al., 1989; Ram et al., 1988) and yields decrease
without any scope for recovery. Data on panicle emergence, and spikelet sterility,
thus provide an indication of timing and intensity of drought (Setter et al., 1993).
Drought is often associated with high temperature and intense solar radiation
alongwith low relative humidity during the dry spells in the rainy season. Soil cracks
develop and strength and bulk density increase, restricting the availability and
uptake of water and nutrients by the plants. Sizeable reductions in soil water
potential due to drought may increase salt concentrations in the soil solution which
may induce dehydration of tissues leading to cell collapse and death of plants.
51

Table 2.

Soil moisture content (% of dry weight) and leaf water potential (-bar)
in rice genotypes under water deficit condition (Pot expt.)

Stages of stress:

IR-28
(Susceptible)

Genotypes
IRAT-142
(Moderate)

KACHANI
(Tolerant)

LSD (P 0.05)

6.8
19.0
8.9

6.4
20.9
8.3

6.6
21.2
8.1

N.S.
0.60
N.S.

7.9
20.0
9.9

7.4
21.0
9.4

7.9
21.5
9.3

N.S.
0.58
N.S.

7.9
21.5
9.9

7.4
22.8
9.0

7.8
23.0
9.0

N.S.
0.60
N.S.

28.31.3
11.20.9
28.351.1

28.31.3
10.10.95
28.01.2

28.11.0
9.80.72
27.50.95

1. Stress at Tillering (30-DAS)

Before stress
At P.W.P.
After revival (48 h.)
2 Stress at Booting (60-65 DAS)
Before stress
At P.W.P
After revival (48 h.)
3. Stress at Tillering + Booting
Before stress
At P.W.P
After revival (48 h.)
4.

Soil moisture
At field capacity
At PWP
After revival

Weed populations under drought invariably exert a threat to crop stand and
render plant easily accessible to insects and diseases. Hence, data on temperature,
relative humidity and light intensity above and inside the plant canopy, weed, insect
and disease profiles may also be included in the environmental measurement data
set (Table 1). All these environmental variables can effectively be integrated through
simulation models for understanding the response of drought on crop growth and
productivity. Such simulation models have been suggested by a number of workers
which include measurements on soil, climate and plant factors (Penning de Vries et
al., 1989; Woopereis et al., 1993; Singh, 1993).
With the above set of environmental and plant measurements, one can easily
predict the timing, duration and intensity of water deficits. This knowledge may
enable breeders to synthesize a new rice genotype suitable for drought prone rainfed
lowlands.

Flooding
Flooding is a serious constraint to rice yields in rainfed lowland and deepwater
areas due to partial or complete submergence of plants. Partial submergence is
defined as when 40 to 99 per cent shoot is submerged in water. Flooding in rice
fields is of two types : i) Flash or intermittent (ii) Stagnant or prolonged.
Approximately 22 m ha of rice is affected by flooding which includes 15 m ha of
52

flash flood areas of rainfed lowland rice and 5 m ha of deepwater rice (Khush, 1984).
In India, approximately 60% of Kharif rice area is subjected to waterlogging due to
flash and stagnant floods (Balkrishna Rao and Biswas, 1979).
Flash flooding refers to a situation where the rice crop gets submerged due
to sudden increase in floodwater for varying periods, normally not exceeding 10-12
days (Dwivedi and Hille Ris Lambers, 1991; Senadhira, 1992). The period of flooding
is often prolonged in Eastern India. Flash flooding normally occurs near river,
riverlets and streams. The rice area near Ghaghara, Ganges, Rapti, Gandak, Kosi,
Baiterni, Mahanadi, Rishikulya, Brahmaputra, Krishna and Cauvery rivers and in
the tidal wetlands coastal area of India, are more vulnerable to flash flooding (Singh,
1982). The flash flooding occurs mainly during vegetative stage of rice and its severity
depends on the intensity and duration of rainfall, nature of water source and
topography.
Rice, by nature, is tolerant to waterlogging, flash flood and associated anoxia
which makes its cultivation possible under flood prone lowland ecosystem. A
number of morphological, physiological and biochemical features are responsible
for tolerance to submergence which are influenced by genotype, submergence depth
and duration and other environmental conditions. Features, like coleoptile, that
tolerate anoxia (Taylor, 1942) and the ability of the coleoptile (Ohwaki, 1967) and
stem (Vergara et al., 1976) to elongate vigorously when submerged, provide chance
to escape complete inundation or minimize the duration. Other attributes include a
root system that can elongate more rapidly in the presence of ethylene (Konings and
Jackson, 1979), a highly porous morphology that facilitates internal diffusion of
oxygen down the plant (Armstrong, 1971) and a hydrophobic outer surface that
traps a coating of air when leaves are submerged. This forms a gaseous film that
may facilitate basipetal diffusion and mass flow of oxygen from the aerial
environment when some leaf tissues remain above the water surface (Raskin and
Kende, 1984).
Despite these adaptive features, most rice cultivars can not survive complete
submergence for a long time, more so, during early stage of development. The
experiments conducted at NDUAT and elsewhere indicate that 45-day old seedlings
showed better tolerance to submergence (7, 11 and 15 days submergence) than 30
or 15-day old seedlings with good genetic variability (Mazaredo and Vergara, 1982;
Chaturvedi et al., 1993).
Experiments conducted at the consortium site at Masodha indicated that with
similar experimental conditions and crop management, rice yields among genotypes
during 1992 and 1993 wet seasons showed a variation of 0.2 to 2.9 t/ha (Table 3). It
is clear from the data that environmental factors like water depth and duration of
submergence, temperature, light, rainfall and humidity have an important role in
changing the phenology, for example, flowering delayed during 1993. (Figures 3,
and 4). Environmental characterization (Table 1) thus, is of utmost importance for
interpreting the performance and determining the stability of the genotypes across
locations and at the same location over different years.

53

Table 3.

Performance of elite
submerged condition.

Genotypes

lines

50% flowering-DAS*

of rainfed

lowland

ERT*/hill

rice

under partially

Yield t/ha

Yield
Difference

1992

1993

1992

1993

1992

1993

7.5

5.3

7.2

+1.9

NDR-30030

113

122

7.8

NDR-30076

118

121

8.2

7.4

5.2

7.0

+1.8

NDR-30039

115

118

8.4

8.1

4.5

6.0

+1.5

NDR-30023

112

118

9.3

7.8

4.3

6.2

+1.9

NDR-40013

120

127

6.7

6.7

4.8

6.4

+1.6

NDR-40032

114

123

8.4

7.4

4.5

7.4

+2.9

Madhukar (Check)

115

126

5.2

7.0

2.0

1.8

-0.2

SEm

0.66

0.71

0.24

0.14

0.23

0.15

LSD (P0.05)

207

223

0.75

0.43

0.72

0.47

Days after sowing;

**

Earbearing tillers

There were large differences in water depth (Figures 3 and 4), showing lower
water levels (mean depth 5.9-11.0 cm) during July and August in 1992 (active tillering
phase) and deeper water level in 1993 (mean depth 19.8 to 23.9 cm). In contrast,
during 1992 higher water depths (11.0-34.5 cm) were observed at flowering to milk
stages than in 1993 (4.0-26.5 cm). The physiological implication of such variations
needs to be looked into as a key to the effect of environmental factors affecting
growth and yield. Water temperature variations and sunshine hours should also be
taken into consideration.
A number of physiological and biochemical explanations for death and
survival under submergence has been reported which include a possible interaction
between CO2 low O2 and high ethylene concentrations (Setter et al., 1988, 1989).
Low photosynthesis due to low CO2 diffusion leads to lower carbohydrate
assimilation and energy deficiency which restrict nutrient uptake and growth.
Submergence may also result in solute leakage due to low O2 concentration and
affect membrane permeability (Smith and ap Rees, 1979).
The rice plants respond and adapt to various types, timing and duration
of flooding in different ways which include elongation for deepwater rice
varieties and submergence tolerance for lowland varieties exposed to transient
flash floods (Hille RisLambers and Seshu, 1982).
A number of environmental and plant factors are associated with damage due
to flooding namely low light, siltation on the leaves, mechanical damage, solute
leakage, limitation to gas diffusion, accumulation of toxic metabolites inside the
plant, insect-pests and diseases (Creenway and Setter, 1994). Additional adverse
54

Fig. 3.

Variation in daily water depth during 1992-93 wet season crop at C.R.S. Masodha, India.

55

Fig. 4. : Weekly meteorological information during wet crop season (1992 & 1993) at C.R.S. Masodha, India.

effects may be associated with desubmergence and post hypoxic injury (see
Crawford, 1992, for review). These factors seldom act alone since alteration in one
component may lead to a number of other adverse effects which ultimately affects
survival and growth of plants under submerged conditions.
Submergence at seedling stage is most harmful (reducing plant population),
followed by submergence at tillering stage which cuts down the tiller production.
Submergence at boot and flowering stages prevents panicle emergence and increases
spikelet sterility (Pandey et al., 1979). High nitrogen level in the soil accentuates
adverse effects of submergence (Palada and Vergara, 1972). Gas diffusion during
submergence is quite low in water, hence any gas (CO2 ethylene and methane)
which is produced under water increases in concentration, while that O2 consumed
decreases in concentration. Any of these changes may reduce growth, and survival
of rice during flooding.
The composition of floodwater may vary with locations, hence proper
measurement and characterization of flood water is essential in order to assess its
effect on plant growth and yield. During 1993 wet season, a brief environmental
characterization of stagnant water in submerged rice fields was done at crop research
stations, Pagalabhari, (Faizabad) and Ghagahraghat, (Bahraich) (Figures 5, 6 and 7).
A diurnal variation in O 2 concentration of water at different depths was noted
showing a decrease in O2 with increase in water depth (0.01 - 0.4 m). Maximum O 2
in water was observed during mid day possibly due to photosynthesis of algae and
submerged rice plant tissues which gradually declined to the lowest before dawn.
Plant respiration during night might have consumed at least some of the dissolved
oxygen (Figure 5). In deepwater, a gradual decline in dissolved O2 was noted upto
1.4 m depth, beyond which, a sharp decrease occurred leading to zero oxygen at 2
m depth well above the soil surface (Figure 6). The temperature profile through
water depth at 10.00 h during day decreased with depth but it ranged only between
27.5 to 28.0 C in the deepwater at Ghagharaghat.

Fig. 5.

Diurnal variation in O 2 concentration of water in partially submerged rice field at C.R.S.

Pagalabhari, Faizabad, lndia Blackbar on X-axis indicates dark period.

57

Fig. 6.

Oxygen concentration of flood water in deep-water rice field at Ghagharaghat, Bahraich, India.
Maximum water depth was 2.4 m.

Fig. 7.

Temperature of water in a field of partially submerged rice at C.R.S. Pagalabhari, Faizabad,

India. (Black bar on X-axis indicates dark period).

58

Diurnal variation in temperature of shallow water at Pagalabhari showed that

the maximum temperature occurred at midday (1200 h) being higher at the surface
layer, decreasing downwards upto 0.4m depth (soil surface). However, during night,
the temperature of the water in deeper layers was higher than surface layers (Figure
7). The lower temperature of the surface water layers during night have been due
to evaporative cooling. These variations in O2/CO2 concentrations and temperature
in rice fields may induce a number of changes in plant metabolism leading to altered
growth and yield. Genotypes with well equipped systems to combat anoxia may
survive and yield better. The data presented here on O2/CO2 and temperature are
only from one wet season and hence need confirmation at other times and locations
to direct the research strategies for alleviating adverse effects of submergence on
rice.
Genotypes with higher total carbohydrate and starch content prior to
submergence along with lower rate of depletion during submergence showed better
tolerance to submergance (Chaturvedi et al., 1993). Since the flood water conditions
in stagnant and turbulent situations vary, the duration of complete submergence
and water current alongwith its turbidity should be measured in all experiments to
quantify the degree of severity of damage to plants.
In summary, the environmental and plant factors need to be carefully
measured during submergence to allow extrapolation and interpretation of results.
These factors include, flood water turbulence, pH, turbidity, O2 and CO 2
concentration, water depth relative to plant height, levels of non-structural
carbohydrate and toxic metabolites inside the plant (Table 1). Future research must
focus on means of manipulating floodwater O2 during submergence as well as on
quantitative determination of the importance of different O2 concentrations to
growth and survival of rice during submergence . At present there is no published
information on the latter though it is well known that O2 is one of the major factors
limiting growth during waterlogging (Jackson and Drew, 1984) and Submergence of
rice plants (Setter et al., 1989).

Salinity
An estimated 150 m ha of current and potential rice lands in the tropics and
subtropics are affected by salinity (Massoud, 1974). in South and South East Asia
alone about 90 m ha of lands suited to rice production, lie idle largely because
of soil toxicities of which nearly 49 m ha are saline and 12 m ha are alkali
(sodic) in nature (Ponnamperuma and Bandyopadhya, 1980). In general, soil
salinity is caused by presence or by intrusion of sea water or by surface
evaporation of brackish soil water. Salt accumulation increases in the drier
climates and diminishes strongly in equatorial climates without a pronounced
dry season.
Rice is generally reported to be a moderately salt tolerant crop but no rice
variety can withstand high salinity or sodicity throughout its growth cycle. High
content of sodium chloride in soil solution with specific conductance values of
more than 8-10 dS/m are harmful to rice plants and cause as much as 50% yield
59

reduction. In tidal areas, the specific conductance values change from day to day
depending on tidal regimes (De Datta, 1981).
Inland saline sodic soils generally have pH more than 7 alongwith high
amounts of soluble salts. The underground water invariably is brackish. Saline-sodic
soils have high amounts of soluble salts along with high concentrations of carbonate
and bicarbonates of sodium. High sodium replaces calcium from the clay complex
with concurrent precipitation of calcium carbonate which forms an impermeable
(upto 1 m thick) kankar layer. The soil becomes strongly dispersed and hence highly
impermeable to water. This necessitates the careful characterization of coastal and
inland saline and sodic soils in terms of topography, bulk density, hydraulic
conductivity, variations in pH, electrical conductivity (ECe) and chemical properties
at essential and desirable levels. Salt affected soils show a number of problems which
hamper normal growth and development of plants (Ponnamperuma and
Bandyopadhya, 1980).
It is evident from the data presented in Tables 4 and 5 that a large variability
exists in the physico-chemical properties of soil including nutritional status which
should be accurately characterized if the results obtained at one location are likely
to be extrapolated to other locations. Underground water is often brackish at certain
locations and this too needs careful monitoring.
Table 4.

Soil characteristics of experimental site of Main Experiment Station,

Kumarganj, Faizabad, India. Data are mean of four replicates; SEMs
were less than 12% of individual values.

Soil parameter

Soil Depth (cm)

Soil

0-15

15-30

Soil pH (1: 2.5)

Sodic
Normal*

9.5
8.5

9.8
8.7

ECe (dS/m)

Sodic
Normal*

1.4
0.6

1.8
1.0

Exchangeable sodium percentage

Sodic
Normal*

35
15

55
17

Hydraulic conductivity (mm/hr)

Sodic
Normal*

0.8
2.5

0.3
2.0

Cation exchange capacity

(me / 100g)
Texture

Sodic
Normal*

14-16
16-19

13-15
14-17

Sodic
Normal*

silty loam
silty loam

Topography

undulating

60

Normal designates a soil where 10 t/ha pyrite was applied to sodic soil two years earlier,
followed by continuous cropping.

In glycophytes including rice, the effect of salinity on plant growth and

development depends on temperature, humidity and radiation (Levitt, 1980).
Vegetative growth of rice, for instance in NaCl affected soils, is better during wet
season, when plants survive 0.4% (W/V) NaCl as opposed to 0.3% in dry season
(Mercado et al., 1974).
Table 5.

Nutritional status of the soil at Main Experiment Station Kumarganj,

Faizabad (As per Table 4). Data are mean of four replicates; SEMs were
less than 14% of the individual values.

Soil parameters

Soil depth (cm)

015

1530
0.050.2
0.20.3

Organic carbon (%)

Sodic
Normal

0.10.2
0.30.4

Available nitrogen (ppm)

Sodic
Normal

20
25

AvailableP2O 5 (ppm)

Sodic
Normal

28
25

Available K2O (ppm)

Sodic
Normal

300
320

Iron

Sodic
Normal

3.8
9.5

3.7
7.2

Manganese

Sodic
Normal

2.9
3.0

2.3
2.7

Zinc

Sodic
Normal

2.2
5.6

1.2
4.5

CaCO3 (%)

Sodic
Normal

6.0
1.5

5.5
2.0

Thickness of CaCO3
(Kankar) layer

Sodic

present
(2540 cm thick)
rarely present

present

Micronutrient Status (ppm)

(DTPA Extractable)

Underground water quality

Normal

good

The response of genotypes to adverse soil conditions vary

upon a number of environmental factors. The data pertaining
productivity clearly indicate that tolerant rice genotypes show less
and biological yields than susceptible genotypes (Tables 6 and 7).

greatly depending
to survival and
Eduction in grain
A large variability

61

was also noted in dry matter partitioning into grains indicating a reduced CO2
fixation or reduced translocation of carbohydrates under high salt concentrations.
Tolerant genotypes had lower salt injury index (0.11 - 0.29) against 0.48 - 0.71 in
susceptible genotypes (Table 7). A comparison of yield versus trait combinations
(Table 8) and harvest index versus trait combinations (Table 9) indicates that
interaction between biological yield, harvest index and grain fertility plays an
important role in determining yield of the genotype. Genotypes with higher total
biomass and better partitioning may be looked into as the future cultivars for salt
affected rice areas.
Table 6.
Genotype

Total biomass production in salt tolerant and susceptible ricegenotypes.

Biological yield g/hill

% reduction
over normal

Normal Soil
(PH 8.5)

Sodic Soil
(PH 9.7)

Usar-1

73.0

70.2

3.8

CSR-10

56.2

49.7

11.6

TCA88-10-1

70.0

54.1

22.7

NDR-80

66.0

61.5

6.8

Purple

58.0

46.3

20.2

NDRK-5003

94.0

65.7

30.1

Pokkali

63.7

43.3

32.0

Pusa- 516

79.0

42.0

46.8

Rajshree

64.0

33.4

47.8

Kasturi

67.0

28.1

LSD (P0.05)

9.77

6.82

58.1
-

Tolerant

Susceptible

The literature frequently contrasts the central dilemma of ion toxicity versus
water deficit (osmotic effect) in saline conditions (Flowers et al., 1977 for halophytes
and Greenway and Munns, 1980 for non-halophytes). Large quantities of salts are
carried through the transpiration stream to the leaves which eventually leads to their
death. NaCl accumulation in leaves is correlated with reduced photosynthetic
activity and with ultrastructural and metabolic damage (Yeo and Flowers, 1986,1989)
and there is now direct evidence that this is mediated by ayoplastic increase in salt
62

concentrations in the expanded leaves (Flowers et al., 1991). There is ample variability
in salt susceptibility between (Yeo et al., 1990) and within (Yeo el al., 1988) varieties,
differences in vigour accounting for much of the variation in the survival in salinity
(Yeo et al., 1990). Short term initial effects of salinity are primarily due to limitation
of water supply, whereas long term effects are due to accumulation of salt within
expanded leaves (Yeo et al., 1991).
Table 7.

Grain Yield and salt injury index in tolerant and susceptible rice
genotypes.

Genotype

Grain yield g/hill

Harvest
index (%)
at pH 9.7

Salt injury
index

Normal Soil
(pH 8.5)

Sodic Soil
(pH 9.7)

Usar-1

24.7

22.0

31

0.11

CSR-10

26.0

20.7

44

0.20

CSR-11

26.9

22.2

33

0.17

TCA 88-10-1

24.3

20.0

37

0.17

NDR-80

26.1

22.6

37

0.13

Purple

24.7

21.9

43

0.11

NDRK-5003

24.5

19.9

30

0.19

Pokkali

22.1

15.5

34

0.29

Pusa-516

23.5

12.2

23

0.48

Tolerant

Susceptible
Rajshree

34.2

10.1

30

0.71

Kasturi

21.1

8.4

30

0.60

LSD (P0.05)

3.96

3.00

4.55

It is highly significant and important that individual plant variability in

sodium uptake and in survival are negatively correlatcd. The presence of other ions
may be expected to affect the absorption of Na+ salts and the injury produced by
them. It has long been known that Ca2+ antagonizes the injurious effects of Na+ and
this effect is largely due to decrease in cell permeability produced by Ca2+ . In contrast
to decreased cell permeability to monovalent cations, Ca2+ also enhances the active
uptake of some ions, particularly K+ and P ions, the "Viets effect". NaCl, infact, induced
a marked efflux of K from leafsegments of bean and barley, and this efflux was prevented
by addition of 1 meq Ca2+ to the external solution (Marschner and Mix, 1973).
The importance of uptake of Na+ , K+ and Ca2+ by rice plants and partitioning
of these ions into different plant parts has been well emphasized. The observations
63

presented in Table-10 indicate the role of beneficial ions like K+ and Ca2+ in mitigating
the detrimental effects of high internal Na + . Almost all the salt tolerant genotypes
invariably possessed higher K/Na and Ca/Na ratios than susceptible ones. A
number of mechanisms for exclusion and sequestering/compartmentation of ions
for mitigating the adverse effects of salinity/sodicity have been proposed by Flowers
el al. (1977, 1991) and Yeo and Flowers (1984) including exclusion, sequestering of
ions in roots, old leaves, or different cellular compartments or dilution by growth.
Table 8.
Genotype

Yield versus trait combinations in rice genotypes planted under sodic

soil (pH 9.7).
Yield
(g/hill)

Biol. Yield
(g/hill)

HI (%)

Grain
fertility (%)

EBT/hill

Salt
injury
index

NDR-80

22.6

61.5

37

87

12

0.13

Usar-1

22.0

70.1

31

90

16

0.11

CSR-11

22.2

56.5

33

83

15

0.17

Purple

21.9

46.3

43

85

18

0.11

CSR-10

21.7

49.7

44

74

16

0.16

CSR-18

21.8

50.0

44

90

15

0.19

IET-12856

22.8

55.1

40

90

12

0.17

IET-12865

21.0

53.0

39

82

12

LSD (P0.05)

N.S.

3.10

4.66

8.15

4.95

0.19
-

Table 9.

Harvest index versus trait combinations in rice genotypes planted

under sodic soil (pH 9.7).

Genotype

HI (%)

Yield
(g/hill)

Biol Yield
(g/hill)

Grain
fertility (%)

EBT/hill

Salt
injury
index

Usar-1

31

22.0

70.1

90

16

0.11

NDRK-5003

30

19.9

65.7

88

12

0.18

NDRK-5002

30

17.1

56.5

79

11

0.37

Rajshree

30

10.1

28.1

46

11

0.70

Kasturi

30

8.4

38.2

43

10

0.60

IET-13408

30

15.9

54.0

91

10

0.47

IET-11353

30

14.3

47.2

80

11

0.25

N.S.

3.68

2.85

6.67

2.71

LSD (P0.05)

64

Table 10.

Genotype

K/Na and Ca/Na ratios in salt tolerant and susceptible rainfed lowland
rice genotypes.

Leaf

K/Na ratio
Stem

Root

Leaf

Ca / Na ratio
Stem

Root

Tolerant
Purple

2.4

1.9

1.1

3.0

4.1

2.1

CSR-10

3.0

3.0

3.1

2.0

2.5

2.2

CSR-5

2.5

3.1

2.5

2.1

2.5

2.2

CSR-18

0.8

1.6

2.1

2.6

Pokkali

1.5

1.7

1.7

1.3

1.8

1.7

TCA-88-10-1

1.9

1.6

1.5

1.3

2.8

1.7

IET-11351

16

1.2

1.2

2.5

1.8

1.6

Sajoo-52

1.6

1.4

2.9

2.2

Pusa-516

1.1

0.9

0.7

0.9

1.0

0.9

Basmati-370

1.0

0.8

1.3

1.1

IET-13407

0.9

0.7

0.9

1.4

1.4

1.1

Susceptible

A distinction should be made between saline soils which contain enough Ca2+
to meet the nutritional requirements of the plants and sodic soils, in which very little
available Ca2+ is present while the concentration of exchangeable sodium is greater
than 15% and ample carbonate and bicarbonate ions are present leading to high pH
of the soil solution. In salt affected soils, accumulation of Na+ occurs in the stem
and leaves of the plants which vary with plant species. Rice accumulated more Na
and Ca than barley or wheat (George, 1967).
We have also observed a selective uptake and accumulation of Na+ , K+ and
Ca+ by different rice genotypes. Tolerant genotypes usually showed higher K/Na
and Ca/Na ratios in leaf, stem and roots than susceptible genotypes. These
measurements may be used as selection criteria for screening rice genotypes for salt
tolerance and should be included for environmental characterization as plant
measurements (Table 1). Importance of vigour in salt tolerance of rice was
emphasized which reduced actual ion concentration due to dilution and thus
decreasing the adverse effects of salinity. Vigour parameters included for
measurement are given in Table 1.
The mechanism of injury or tolerance to coastal and inland salinities may vary,
as in the former case, the plants are exposed to intermittent high levels of salinity
due to tidal inundations which often fluctuates. Inland salinity, though also fluctuates
with other environmental factors like rainfall, temperature, humidity and light, but
plants are exposed to a continuous stress, which enables them to adjust against
slowly increasing ion concentrations within tissues. In both cases the slow build up
65

of ions to inhibitory concentrations may explain why growth reductions over short
exposure periods of salinity are less meaningful than growth or survival measured
over several weeks (Yeo and Flowers, 1984; Akita and Cabuslay, 1986).
In rice growing areas, the inland saline soils especially in India and Pakistan
are notorious due to associated alkalinity problems alongwith nutritional
imbalances. Rice plants grown in saline-alkali or alkali soils are exposed to high pH,
low to very low hydraulic conductivity, deficiency of nitrogen, zinc and iron, besides
high soluble salts. These factors create an array of complex situations for plants
growing therein. Neue et al. (1990), found that salinity tolerance scores for rice are
significantly correlated to other stresses such as alkalinity and peatiness, P deficiency
to alkalinity and zinc deficiency and Zn deficiency to Fe toxicity. Pokkali, one of
the tolerant rice varieties to salinity could not match well to CSR-10, NDR-501 and
TCA-88-10-1 in our rice screening programme for rainfed lowland salinity/alkalinity
(Tables 6, 7). The reason is clear on the basis of the soil characterization since most
of the saline soils are also highly alkaline (pH upto 10.5) and deficient in nitrogen,
zinc and iron. This result contradicts the generalization that more vigorous and tall
genotypes are tolerant to salinity (see also Tables 6 and 7). Dilution effect may be
true over short term and needs to be evaluated further in long term experiments.
The complexities of saline environments, thus, render it difficult to
characterize. The EC of the soil solution is the most important component to quantify
the overall ions, and needs a careful monitoring during the experiment. The soil
matric potential, water content, and nutrient status may also be given careful
consideration. Soil pH, hydraulic conductivity, and the nature of CaCO3 hardpan
below root zone may essentially be measured in order to seggregate the plant
responses to saline versus alkaline environments.

CONCLUSION
The forgoing discussions clearly indicate the importance of environmental
characterization for drought, flood and salinity situations having enormous amount
of variability across locations and seasons. Such measurements enable extrapolation
of findings from one site to the other sites or locations in the rice growing countries.

REFERENCES
Akita S, Cabuslay C S (1986) Physiology of varietal response to salinity. II. Response to different salt
concentrations. Jap. J. Crop Sci. 55 : Extra Issue 1:28-29.
Armstrong W (197l) Radial oxygen losses from intact rice roots as affected by distance from the apex,
respiration and waterlogging. Physiol. Plant. 25: 192-197.
Austin R B, Margin R B, Ford M A, Blackwell R D (1980) Contribution to grain yield from pre-anthesis
assimilation in tall and dwarf genotypes in two contrasting seasons. Ann. Bot.. 45 : 309-319.
Balkrishna Rao, Biswas S (1979) Rainfed lowland rice in India(Pages 87-94). In : Rainfed lowland rice;
Selected papers from 1978 International Rice Research Conference IRRI, Los Baos, Philippines.

66

Castillo E G, Buresh R J, Ingram KT (1992) Lowland rice yield as affected by timing of water deficit and
nitrogen fertilizer. Agron. J. 84 : 152-158.
Chaturvedi G S, lngram KT (1988) Assimilate translocation of lowland rice in response to water deficit
and shade. Paper presented at FCSSP Annual Meeting held at Davao City 27-30 April, 1988.
Chaturvedi G S, Misra C H, Singh O N, Pandey C B, Yadav V F, Singh A K, Dwivedi J L, Singh B B, Singh
R K (1995) Physiological basis and screening for tolerance for flash flooding. Pages 79-96. In : Rainfed
Lowland Rice - agricultural research for high risk environments. KT Ingram, ed., International Rice
Research Institute, Los Banos, Philippines.
Crawford R M M (1992) Oxygen availability as an ecological limit to plant distribution. Adv. in Ed. Res.
23: 93-185.
De Datta S K (1981) Principles and practices of rice production. John Wiley and Sons, New York, USA.
Dwivedi J L, Hille RisLambers, D (1991) Deepwater rices varietal improvement. IRRI,
Philippines.

Los Banos,

Ekanayake I J, De Datta S K, Steponkus P L (1989) Spikelet sterility and flowering response of rice to
water stress at anthesis. Annals Bot. 63 : 252-264.
Flowers T J, Hajibagheri M A, Yeo A R (1991) Ion accumulation in cell walls of rice plants growing in
saline condition: evidence for the Oertli Hypothesis. Plant cell Environ. 14 : 319-325.
Flowers T J, Trake P F, Yeo A R (1977) The mechanism of salt tolerance in halophytes. Annu. Rev. Plant.
Physiol. 28 : 89-121.
Garrity D P, Oldeman L R, Morris R A, Lenka D (1986) Rainfed lowland rice ecosystems : Characterization
and distribution. Pages 3-240. In : Progress in rainfed lowland rice IRRI, Los Banos, Philippines.
George L Y (1967) Accumulation of sodium and calcium by seedlings of some cereal crops under saline
conditions. Agron. J. 59 : 297-299.
Greenway H, Munns R (1980) Mechanism of salt tolerance in non - halophytes Annu. Rev. Plant Physiol.
31: 149-190.
Greenway H, Setter T L (1994) : Is there anaerobic metabolism in submerged rice plants? A View point
(This proceeding).
Herdt R W, Barker R (1977) Multi-site test environments and breeding strategies for new rice technology.
IRRI Res. Pap. Ser, 7. IRRI, Los Banos, Philippines, 32p.
Hille RisLambers D, Seshu D V (1982) Some ideas on breeding procedures and requirements for deepwater
rice improvement. Pages 29-34. In.: Proceedings of 1981 International deepwater rice workshop. IRRI, Los
Banos, Philippines.
Hsiao T C(1982) The soil-plant atmosphere continuum in relation to drought and crop production. Pages.
39-52. In: Drought resistance in crops with emphasis on rice. IRRI, Los Banos, Philippines.
Jackson M B, Drew M C (1984) Effects of flooding on growth and metabolism of herbaceous plants. Pages
47-128. In : Flooding and plant growth. T.T. Kozlowski ed. Academic Press, Orlando, Florida.
Jackson M L (1973) Soil Chemical analysis. Prentice Hall of India Pvt. Ltd., New Delhi.
Khush G S (1984) Terminology for rice growing environments. IRRI, Los Banos, Philippines.
Konings H, Jackson M B (1979) A relationship between rates of ethylene production by roots and the
promoting or inhibiting effects of exogenous ethylene and water on root elongation. Zeitschrift fur
Pflanzenphysiologie 92: 385-397.
Levitt J (1980) Responses of plants to environmental stresses Vol. II. Water, radiation, salt and other stresses.
Academic Press, New York, London, Toronto.

67

Marschner H, Mix G (1973) Effect of sodium chloride and mycostatin on the mineral content of leaf
segments and on the fine structure of chloroplasts. Z. Pflanzenermechi Bokenkd 136: 203-219.
Massoud F I (1974) Salinity and alkalinity as soil degradation hazards, FAO/UNEP expert consultation
on soil degradation. FAO, Rome.
Mazaredo A M, Vergara B S (1982) Physiological differences in rice varieties tolerant of and susceptible
to complete submergence. Pages 327-341. In. : Proceedings of the 1981 International deepwater rice
workshop IRRI, Los Banos, Philippines.
Mercado B T, Malabayabas C A, Gumasing S R (1974) Growth of some rice varieties in NaCl salinated
soils as affected by the season. Philip. Agric. 58: 40-49.
Neue H U, Lantin R S, Cayton M T C, Autor N U (1990) Screening of rices for adverse soils tolerance.
Pages 523-531. In: Genetic aspects of plant mineral nutrition. N.EI. Bassam et al., eds. Kluwer Academic
Publ, The Netherlands.
Ohwaki Y (1967) Growth of rice coleoptiles in relation to oxygen concentrations. Scientific reports of
Johoku University, Series IV Biology 33 :I-5.
OToole J C, Baldia E (1982) Water deficits and mineral uptake in rice. Crop Sci. 22: 1144-1150.
OToole J C, Moya T B (1978) Genotypic variation in maintenance of leaf water potential in rice. Crop Sci.
18: 873-876.
Palada M C, Vergara B S (1972) Environmental effects on the resistance of rice seedlings to complete
submergence. Crop Sci. 12:209-212.
Pandey H K, Mitra B N, Ghosh B C (1979) Flood susceptibility of semi dwarf rices and their suitability
for lowlying areas. Pages 185-195. In: Proceedings of 1978 International deepwater rice workshop. IRRI,
Los Banos, Philippines.
Penning de Vries F W T, Jansen D M, Ten Barge H F M, Bakema A (1989) Simulation of ecophysical
processes of growth in several annual crops. IRRI, Los Baos, Philippines.
Ponnamperuma F N, Bandyopadhya A K (1980) Soil salinity as a constraint on food production in the
humid tropics. Pages 203-216. In: Soil related constraints to food production in the tropics. IRRI, Los
Banos, Philippines.
Ram P, Ram P C, Singh B B (1988) Response of rice genotypes to water stress imposed at the tillering and
boot stages of growth. lndian J. Plant Physiol. 31: 308-311.
Raskin I, Kende H (1984) Regulation of growth in stem sections of deepwater rice. Planta 160:66-72.
Reyniers F N, Troung - Binh, Jacquinot L, Nicou R (1982) Breeding for drought resistance in dry land rice.
Pages 273-292. In: Drought resistance in crops with emphasis on rice. IRRI, Los Banos, Philippines.
Senadhira D (1992) Abiotic stresses of rice, IRRI Rice Research Seminar, 30 January 1992 IRRI, LOS Banos,
Philippines.
Setter T L, Kupkanchanakul T, Kupkanchanakul K, Bhekasut P, Wiengweera A, Greenway H (1988)
Environmental factors in deepwater rice areas in Thailand : O2, CO2 and ethylene. Pages 55- 66. In:
Symposium of the 1987 International deepwater rice workshop. IRRI, Los Banos, Philippines.
Setter T L, Greenway H, Kupkanchanakul T (1989). Submergence of rice II. Adverse effect of low CO2
concentrations. Aust. J. Plant Physiol. 16: 265-278.
Setter TL, Ingram K T, Tuong TP (1993) Environmental characterization requirements for strategic research
in rice grown under adverse conditions of drought, flooding or salinity. Pages 3-18. In : Rainfed lowland
rice- agricultural research for high-risk environments. KT Ingram, ed, IRRI, Los Banos, Philippines.
Setter T L, Ramkrishnayya G, Ram Maurya P C, Singh B B (1995) Environmental characteristics of
floodwater in eastern India; Relevance to flooding tolerance of rice. lndian J. Plant Physiol. 38 : 34-40.

68

ShardaKov V S (1948) New field method for the determination of the suction pressure of plants. Dold.
Akad. Nauk SSSR 60 : 169172.
Singh U S (1982) Deepwater rice in India : problems and prospects Pages. 355362 In: Proceedings of 1981
International deepwater rice workshop. IRRI, Los Baos, Philippines.
Singh U, (1993) Systems approach to understanding and predicting nutrient response in rainfed lowland
rice. Paper presented at Rainfed lowland Programme Meeting, 19-20 January. IRRI, Los Baos, Philippines.
Smith A M, ap Rees T (1979) Effects of anaerobiosis on carbohydrate oxidation by roots of Pisum sativum.
Phytochemistry 18: 14531458.
Taylor D L (1942) Influence of oxygen tension on respiration, fermentation, and growth of wheat and rice.
Amer. J. of Bot. 29: 721738.
Vergara B S, Jackson B, De Datta S K (1976) Deepwater rice and its responses to deepwater stress. pages
301-320. In: Climate and Rice, IRRI, Los Baos,. Philippines.
Woopereis M C S, Kropff, M J, Bouma, J (1993). Soil data needs for systems approaches at different levels
of detail and application of gas to predict water limited crop production. In Research programme on soil
management for increased and sustainable rice production. Woopereis et.al. eds. IRRI, Los Baos,
Philippines.
Yambao E B, Ingram K T (1988) Drought stress index for rice. Phil. J. Crop Sci. 13 : 105111.
Yeo A R, Flowers T J (1984) Mechanisms of salinity resistance in rice and their role as physiological criteria
in plant breeding. Pages 151170. In: Salinity tolerance in plants-strategies for crop improvement, R.C.
Staples and G.A. Toenniessen, Eds. Wiley International, New York, USA.
Yeo A R, Flowers T J (1986) The physiology of salinity resistance in rice Oryza
(
sativa L.) and a pyramiding
approach to breeding varieties for saline soils. Australian J. Plant Physiology 13: 7591.
Yeo A R, Flowers T J (1989) Selection for physiological characters- examples from breeding for salt
resistance. Pages 217234. In: Plants under stress, HG. Jones, T.J. flowers and M.B. Jones eds. Cambridge
University Press, Cambridge, Great Britain.
Yeo A R, Yeo M E, Flowers T J (1988) Selection of lines with high and low sodium transport from within
varieties of an inbreeding species, rice ( Oryza sativa L.) New Phytologist 110 : 1319.
Yeo A R, Yeo M E, Flowers T J, Flowers S A (1990) Screening of rice Oryza
(
sativa L) genotypes for
physiological characters contributing to salinity resistance and their relationship to overall performance.
Theoretical and Applied Genetics 79: 377384.
Yeo A R, Lee K S, Izard P, Boursier P J, Flowers T J (1991) Short and longterm effects of salinity on leaf
growth in rice (Oryza sativa L.) J. Exp. Bot. 42 : 881889.

69

NUTRIENT UPTAKE BY ROOTS OF CEREALS IN

OXYGEN DEFICIENT ENVIRONMENT
M .C. Drewl

SUMMARY
Uptake and transport of nutrient ions by roots of higher plants is strongly
dependent on oxidative phosphorylation in order to supply adequate ATP. When
aerobic respiration is slowed down by O2- deficiency in the rooting zone, roots
become hypoxic or anoxic; mineral nutrition may be affected. In barley and wheat,
phloem-mobile nutrients are reallocated from older, senescing leaves to the younger
and growing ones; but the mineral nutrition of the plant as a whole is not sufficient
and deficiency symptoms appear. Under saline conditions, Na+ fails to be excluded
by roots of maize and reaches leaves in injurious amounts. Uptake of Cl- from low
concentration, a thermodynamically active process is strongly inhibited by anoxia,
even when roots are not totally depleted of ATP.
When cereals such as barley, wheat or maize develop aerenchymatous roots
in response to O2 deficiency, nutrient uptake can continue despite the loss of
many cells in the root cortex. Internal transport of O2 from the leaves improves
the energy status of the root. In rice, and other aquatic plants that form an
extensive aerenchyma, suberization and lignification of the root axes assists in
O2 conservation by minimizing radial leakage and loss to the rhizosphere.
However, these structures sometimes act as barriers to radial, inward, transport
of nutrient ions in other species, so that uptake probably depends largely on the
formation of fine laterals that lack structural barriers.

INTRODUCTION
Most cereal crops are not well-adapted to wetland conditions, showing a
variety of responses to a sudden excess of water in the rooting zone, including
inhibition of leaf extension and tillering, precocious senescence of older leaves,
stomatal closure and reduction of net photosynthesis, and finally in a drop in
economic yield (Van't Woudt and Hagan 1957; Cannell and Jackson, 1981; Meyer et
al., 1985; Hodgson et al., 1990). Such flooding symptoms are undoubtedly a
consequence of the inadequate performance of the roots. In transiently flooded soil,
dissolved O2 is quickly depleted by the rapid respiration of roots and
micro-oganisms when temperatures are high, (Meyer et al., 1985; Hodgson et al., 1990)
1 Department of Horticultural Sciences, Texas A & M University, College Station, Texas 77843-2133, USA.

so that there is insufficient O2 to maintain oxidative phosphorylation in the roots

(Pradet and Bomsel, 1978).
Under these conditions the soil becomes reducing and there is an accumulation
of metal ions, organic acids and volatile substances, including CO 2 and ethylene,
that are injurious to plants, or exert distinct physiological responses
(Ponnamperuma, 1972, 1984; Drew and Lynch, 1980). However, potential
phytotoxins do not appear immediately in the soil solution following the onset of
flooding, but rather accumulate over an extended period of time before they reach
active concentrations. During that time, roots are exposed to O 2 deficient conditions
which alone are sufficient to bring about many of the typical symptoms of flooding
injury (Trought and Drew, 1980c).
This review briefly examines the overall changes in mineral nutrition in cereals
that occur with flooding, and considers the mechanisms by which ion transport in
roots is modified under conditions of restricted O2 supply in the field and in the
laboratory condition.
OXYGEN DEFICIENCY AND MINERAL NUTRITION
Transient flooding under field condition can result in a complex situation with
respect to O2 supply to roots. During the growing season, depending on soil
temperature and respiration rate, dissolved O 2 in the rooting zone gradually declines
towards zero over a period of hours to days. Within the roots particularly in the
apical zones where respiration rates are highest, O2 concentration gradients may
develop more rapidly than in the surrounding soil (Armstrong and Beckett, 1987;
De Willigen and Van Noordwick, 1989), and cells become hypoxic; that is, oxidative
phosphorylation is slowed. Cells interior to the endodermis, with less access to the
diminishing O2 concentration in the soil, quickly becomes anoxic, that is, oxidative
phosphorylation becomes negligible relative to ATP production in fermentation
(Pradet and Bomsel, 1978). Because of the very low Km of a cytochrome oxidase for
O2 such anoxic cells are virtually anaerobic, or O2 -free. At the same time, for older
root zones near the soil surface and close to the junction with the shoot, sufficient
O2 may enter by diffusion from the superficial, more oxygenated layer of soil, or
through intercellular spaces (Saglio et al., 1983). Such root zones may continue their
physiological functions, although they would represent only a small fraction of the
root system for a mature plant. In the field, further heterogeneity arises from the
more rapid depletion of O2 from within soil aggregates compared with the larger,
drained pores between aggregates and peds (Greenwoood and Berry, 1962), or from
pockets of O2 trapped within the soil, such as occurs with a perched water table.
Laboratory experiments cannot readily reproduce such situations.
Field observations, as well as those of plants in undisturbed soil in lysimeters
(Belford, 1981) or in columns of remixed soil, (Leyshon and Sheard 1974; Drew and
Sisworo 1979; Trought and Drew, 1980a), have all shown distinct changes-in mineral
nutrition of the shoot when the O 2 supply is restricted (reviewed by Drew, 1988,
1991; Drew and Lynch, 1980). Typically, N, E K, Ca and Mg all decrease in
concentration in the shoots while Fe, Mn Na and Cu increase. What changes in the
71

soil environment account for this? We investigated this question with barley (Drew
and Sisworo, 1979) and wheat (Trought and Drew 1980a) grown in 7.5 cm diameter
cylinders of a sandy loam soil (2% organic matter). After barley grew for 13 days
(18C day 16 C night) in a controlled environment the soil was water saturated and
the changes in the composition of the soil solution as well as various plant
parameters, were measured with time. The most significant change in the soil was
the loss of dissolved O2 , which was essentially depleted in 24 h period. The initial
symptoms of flooding damage to barley shoots, such as slow increase in fresh weight,
inhibition of leaf extension, and chlorosis of the tops of the oldest (lower) leaves
appeared in 2-4 days. After only 2 days, there was a marked decrease in concentration
of N, P and K in shoots. At that stage, none of the potential toxins that accumulate,
as the soil redox declines, had reached appreciable concentrations and there was still
ample free nitrate. It was concluded that the early responses to flooding, including
the lowering of mineral nutrient status, was the result at least in part of the inhibition
of ion transport by roots through lack of O2. The premature chlorosis or senescence
of older leaves with flooding was accompanied by a rapid loss of N from those
leaves, indicating a remobilization from older to younger leaves. With wheat in soil
columns (Trought and Drew 1980b), flooding strongly inhibited net uptake of N, P
and K into the shoots. Remobilization from older to younger leaves was found to
include all three of these phloem - mobile nutrients.
Increased concentrations of Fe and Mn in shoots in the longer term accord
with the much greater solubility in the soil water at low redox potentials, while Na +
exclusion by roots fails under O2 -deficient conditions. If O2 -deficiency itself accounts
for the early stages in flooding response in cereals, it seemed reasonable to expect
that plants in de-oxygenated nutrient solution would respond similarly to those in
flooded soil. Wheat was grown in nutrient solution for 13 days at 14C in a controlled
environment room, and then the mots were made suddenly O2 deficient by sparging
with N2 gas (Trought and Drew, 1980c). Essentially all of the responses recorded
with soil flooding occurred with plants in deoxygenated nutrient solution, where
none of the potentially harmful products of low soil redox could accumulate.
Net uptake of nutrients into the shoot was almost eliminated with
deoxygenation of the nutrient solution. The average concentration of nutrients in
the xylem sap was calculated from measurements of the volume of water transpired
and the net transfer of nutrients to leaves; for P, K, Ca and Mg these concentrations
were equal to, or less than, those in the nutrient solution, indicating that mass flow
of water through the roots to leaves could account for the small continuing transfer
of nutrients. For N, the calculated concentration in the xylem sap exceeded that of
NO3 in the nutrient solution. It may be suggested that N might be remobilized from
within the senescing root system, or be absorbed by the small number of
aerenchymatous roots that were beginning to form at the shoot base and enter the
nutrient solution toward the end of the experimental period.
If mineral deficiency of shoots is a contributory factor in flooding damage,
might it be possible to overcome some of the symptoms by delivering additional
nutrients to the shoots? It has long been recognized that under outdoor conditions,
72

heavy dressings of N to the soil surface can help alleviate the detrimental effects of
soil flooding (Van Hoorn, 1958; Belford, 1981). There is a variety of explanations for
this, including the delaying effect of NO3- on the lowering of the soil redox potential
(Ponnamperuma, 1972) and replacement of NO -3 lost in denitrification (Gambrell
and Patrick, 1978). However, with wheat we used a split root technique to show that
as long as a small part of the root system was kept oxygenated and supplied with
a full complement of nutrients, nutrient uptake and growth were comparable to that
of aerobic control plants (Trought and Drew, 1981). A remarkable alleviation of
flooding injury in soil columns was obtained with barley, simply by regular additions
of Ca (NO3)2 solution to the soil surface (Drew et al., 1979). Here, presumably the
hypoxic roots at the surface absorbed nutrients sufficiently rapidly to compensate for
the disfunction of the remainder in the deeper soil. It seem as if mineral nutrition
together with its interaction with the plants hormonal status can largely overcome the
deleterious effects on the shoot.
It has been suggested that NO3- can act as an alternative electron acceptor to
O2 under anoxia, and thereby, permit regeneration of NAD from NADH to maintain
glycolysis(Garcia-Novo and Crawford, 1973). In effect, a competition was envisaged
between nitrate reductase and alcohol dehydrogenase (for NADH) such that
anaerobic production of ethanol would be inhibited. However, experiments with
rice roots (Reggiani et al., 1985) failed to confirm this expectation. Moreover, with
maize roots no evidence could be found that the presence of NO -3 enhanced the
energy status under strict anoxia (Saglio et al., 1988); a more rapid provision of NAD
might be expected to accelerate glycolysis and substrate- linked phosphorylations
to yield ATP. Thus the effectiveness of 'nitrate respiration' in roots of higher plants
remains in doubt.
ANOXIA AND ION TRANSPORT BY NON-AERENCHYMATOUS ROOTS
Absorption of cations and anions across the plasma membrane of root cells
is linked directly or indirectly to the activity of H+ -translocating ATPases that pump
protons out of the cell towards the external medium (Poole 1978; Serrano 1989;
Sanders, 1990). The inward (passive) flux of protons, in co-transport with other ions
is the initial step in the absorption of nutrient ions into the plant. After radial
transport across the root, transport from xylem parenchyma into the xylem likewise
depends on the activity of a similar H+ -pump (Hanson, 1978; Clarkson et al., 1984).
Higher plant cells contain no reserves of ATP, which in metabolically active
cells like those of the root apical zone, is sufficient to maintain metabolism for less
than 60 seconds at room temperature (Roberts et al., 1985). The immediate effect of
sudden imposition of anoxia is thus to deplete cells of nucleotide triphosphates, and
thereby arrest the activity of proton pumps. Depolarization of membrane potential
is therefore an almost immediate response to the sudden imposition of anoxia
(Cheeseman and Hanson 1979; Buwalda et al., 1988), and presumably the dissipation
of the H+ gradient across the plasma membrane curtails active ion transport (Rao
and Rains, 1976; Cheeseman and Hamon, 1979; Jacoby and Rudich, 1980). In anoxic
maize roots, passive inward movement of K+ could still continue because of a-100mv
73

diffusion potential at the membrane (Cheeseman and Hanson, 1979), indicating that
a general degeneration of membrane properties did not take place in the short term.
This accords with other observations that the apical zone of maize mots, the zone
most sensitive to anoxia, retains viability for 15-18 hours when anoxically shocked
(Roberts et al., 1984; Johnson et al., 1989). However, the evidence suggests that
anaerobic generation of ATP during fermentation is insufficient to energize ion
transport in competition with all the other requirements for ATP in energy-depleted
cells.
With long term anoxia, a seemingly generalized decline in plasma membrane
properties takes place with loss of inorganic ions and organic solutes from roots of
wheat (Greenway et al., 1992). The concentrations of solutes remaining in the roots
at 70 hours of anoxia, as a percentage of the initial values, were K+5%, free amino
acids 41%, soluble sugars 7%. However, in experiments with roots that had been
aciclimated by exposure to solutions bubbled with about 6% O2, losses at 20 hours
of anoxia were measured for different mot zones. For the apical 1mm zone for those
mot tips that had lost their elongation potential (i.e. were moribund), there was little
retention of K+, amino acids and sugars. Root tips that could still grow, and older
mot zones (1-5mm and 10-20 mm) maintained similar concentrations of solutes as
the controls. There was a close association between generalized solute leakage and
loss of cellular viability of the root tips of wheat.
Subapical zones of hypoxically acclimated roots of wheat retain their ability
to transport ions after 24 h of anoxia, when re- exposed to air (Greenway et al., 1992).
Evidently in these relatively anoxia-tolerant cells, which do not leak readily like
anoxically shocked ones, the machinery for ion transport is not lost.
The effect of O2-deficiency on ion transport to the xylem has long been
a topic of interest. The hypothesis of Crafts and Broyer (1938) concerning the
mechanism for ion transfer from xylem parenchyma cells to the xylem
postulated that those cells bordering the xylem were leaky as a result of a
deficiency of O2 within the stele. Under most conditions of aeration, no such
deficiency of O2 occurs, and the 2-pump hypothesis is now widely accepted.
However, when aeration becomes less than ideal, there is evidence that the
stele in maize roots does become anaerobic (Thomson and Greenway, 1991) :
for enzymes like pyruvate decarboxylase and alcohol dehydrogenase that are
indicative of anaerobic, fermenting cells, there is induction in the stele but not
the outer, more aerobic cortex.
OXYGEN DEFICIENCY AND SALINITY
Flooding of the soil with saline water represents a special situation of economic
interest, because irrigated agriculture in arid and semi-arid climates is often
associated with poor soil drainage and excess salts (Carter, 1975). In terms of root
function, environments low in O2 allow a greater transfer of Na+ from roots to shoots.
With adequate oxygenation, Na+ is actively pumped out of the cytoplasm of root
cells to the outer solution, and also sequestered by xylem parenchyma cells (Shone
et al., 1969), thereby depleting Na+ in the ascending xylem sap. Such mechanisms
74

are dependent on ATP and break down when deprived of O2 (Drew and Lauchli,
1985; Drew et al., 1988). The additional flux of Na+ to the xylem, together with
inhibition of K+ transport, can lead to extreme changes in the ratio Na : K entering
the shoots. In one study by Drew et al. (1988) this ratio increased by a factor of 860
for O2 deficient roots when the concentration of NaCl in the external solution
increased from 1.0 to 50.0 mM. Absorption of C1 into roots is normally dependent
on ATP for active transport, and O 2-deficiency strongly inhibits uptake at low to
moderate concentrations (upto about 100mM) in maize (Ghosh and Drew,
unpublished). However, at higher concentrations we found that passive transport
was appreciable and perhaps as a result of depolarization of plasma membranes,
C1 translocation to the shoot was enhanced relative to aerobic controls. For salt
sensitive species like rice and maize (Yeo et al., 1988), root O2 -deficiency can thus
greatly exacerbate the detrimental effects of excess salts.
NUTRIENT UPTAKE BY AERENCHYMATOUS ROOTS
The internal transfer of O2 from shoots to roots within aerenchymatous cortical
tissue is an important mechanism for complete or partial avoidance of O2 deficiency
in many cereal species. In rice, the system of interconnected gas spaces or lacunae
is very well developed, as it is in many aquatics, so that roots can extend considerable
distances into anaerobic soil (Armstrong, 1979), but the demand for O2 is just as
great as in non wetland species (Armstrong and Webb, 1985). The low porosity to
O2 of the hypodermis, other than in the apical zone, ensures rice roots to avoid loss
of O 2 to the surrounding soil, but rather conserve it for respiration and for
oxygenation of the rhizosphere in the tip zone. Additionally, the extension of
prominent intercellular spaces into the apical meristem probably assists diffusion of
gaseous O2 to a group of cells with a very high respiratory requirement for O2
(Armstrong, 1979). Non-wetland species are not so well-adapted, failing to form an
impervious hypodermis, or such a prominent aerenchyma, or intercellular spaces in
the meristem.
In non-wetland cereals, aerenchyma formation in the roots requires induction by
hypoxia (Thomson et al., 1990). Such conditions enhance ethylene biosynthesis and
entrapment within root tissues, the ethylene in turn promoting lysis of cells in the cortex
to form interconnected lacunae (Drew et al., 1979; Jackson et al,, 1985a). Aerenchyma
formation in rice is mostly constitutive, failing to demonstrate a clear signaling by
ethylene (Jackson et al., 1985 b).
What is the significance of such changes in root anatomy to radial ion transport
to the xylem? In aerenchymatous roots of maize, radial transport of K + and Pi under
well oxygenated conditions was not inhibited by the degeneration of many of the
cortical cells Drew et al, 1980; Drew and Saker 1986). The remaining radial files of
intact cortical cells, and radial cell wall remnants, appeared able to transport ions
across the cortex without lowering the overall rate of transport (Drew and Fourcy,
1986). In situ, such aerenchymatous roots are sufficienctly well supplied with O 2 to
maintain energy metabolism (Drew et al., 1985), so that it seems reasonable to
suppose that ion transport can continue.
75

In rice and other species with strongly suberized and lignified hypodermal
layers (Clark and Harris, 1981), the question arises as to whether these act as barriers
to ion transport (reviewed by Drew, 1987). In the sand sedge (Carex arenaria), the
hypodermis becomes suberized within 2 to 5 cm of the root tip and becomes highly
impermeable to water and inorganic nutrients (Robards et al., 1979), so that uptake
is confined to the finer lateral roots lacking such anatomical features. In maize, the
permeability of the hypodermis depends on the humidity to which the mot has been
exposed, with greater impermeability resulting from exposure to air (Ferguson and
Clarkson; 1976; Peterson and Perumalla, 1984; Clarkson et al., 1987). In sorghum
exposed to dry soil, lignification and suberization of the hypodermis and endodermis
occur unusually close to the mot tip, and the late maturing metaxylem vessels fail
to lose their cross walls (Cruz et al., 1992). The combined effect is to make such roots
ineffective in water absorption. Relatively little work has been published on the
precise zones along rice seminal and nodal roots that contribute to ion transport, so
that the consequences of anatomical changes are less well understood. Such work
is best done with radio-labeled tracers (Peterson, 1987), because although apoplastic
fluorescent dyes provide useful evidence concerning changes in the resistance of
potential pathways for flow of solutes and water, their larger molecular size is likely
to become excluded where some passage of smaller molecules can continue.

CONCLUSION
For non-wetland species, oxygenation of the root system can greatly modify
its function in plant mineral nutrition. Simultaneously, changes in permeability to
water, and the synthesis of phytohormones and their transfer to the shoot, contribute
to the symptoms of flooding damage. The ability of root cells to maintain viability
under anoxia, together with the rapid formation of a new, aerenchymatous root
system of a size that can replace the function of the original, damaged one, are
features that assist greatly a plant's tolerance of transient flooding. Better
understanding of the genetic mechanisms of anoxia-tolerance and anoxia-avoidance
will help in further improvement of performance of flood-sensitive dryland cereals.

REREFENCES
Armstrong W (1979) Aeration in higher plants. Adv. Bot. Res. 7:225-331.
Armstrong W, Beckett P M(1987) Internal aeration and the development of stelar anoxia in submerged
roots: A multishelled mathematical model combining axial diffusion of oxygen in the cortex with radial
losses to the stele, the wall layers and the rhizosphere. New Phytol. 105 :221-245.
Armstrong W, Webb T (1985) A critical oxygen pressure for root extension in rice. J. Exp. Bot. 36: 1573-1582.
Belford, R K(1981) Responseof winter wheat to prolonged waterlogging under outdoor conditions. J.Agri.
Sci. Camb. 97: 557-568.
Buwalda F, Thomson C J, Steigner W, Barrett-Lennard E G, Gibbs J, Greenway H (1988) Hypoxia induces
membrane depolarization and potassium loss from wheat roots but does not increase their permeability
to sorbitol. J. Exp. Bot. 39: 1169-1183.

76

Cannell R Q, Jackson M B (1981) Alleviating aeration stresses. Pages 141-192. In: Modifying the root
environment to reduce crop stress. G.F. Arkin and H.M. Taylor, eds. American Society of Agricultural
Engineering, Madison.
Carter D L (1975) Problems of salinity in agriculture. Pages 25- 35. In : Plants in saline environments. A
Poljakoff-Mayber and J. Gale, eds. Springer Verlag, Berlin.
Cheeseman J M, Hanson J B (1979) Energy-linked potassium influx as related to cell potential in corn
roots. Plant Physiology. 64: 842-845.
Clark L H, Harris W H (1981) Observations of the root anatomy of rice ( Oryza sativa L.) Amer. J. Bot.
68:154-161.
Clarkson D T, Robards A W, Stephens J E, Stark M (1987) Suberin lamellae in the hypodermis of maize
( Zea mays ) roots; development and factors affecting the permeability of hypodermal layers. Plant Cell
Environ. 10:83-93.
Crafts A S, Broyer T C (1938) Migration of salts and water into xylem of roots of higher plants. Am. J.Bot.
25:529-535.
Cruz R, Jordan W, Drew M C (1992) Structural change and associated reduction of hydraulic conductivity
in roots of Sorghum bicolor L. following exposure to water deficit. Plant Physiology. 99: 203-212
De Willigen P, Van Noordwick M (1989) Model calculations on the relative importanceof internal longitudinal
diffusion for aeration of roots on non-wetland plants. Plant and Soil 113: 111-119.
Drew M C (1988) Effects of flooding and oxygen deficiency on plant mineral nutrition. Page 115-159. Vol.
111 In : Advances in plant nutrition. A. Lauchli and P.B. Tinker, eds. Praeger, New York.
Drew M C (1987) Function of root tissues in nutrient and water transport. Pages 71-101. In : Root
development and function effects of the physical environment. P J Gregory, J V Lake and A H Rose eds.
Cambridge University Press, Cambridge-London.
Drew M C (1991) Oxygen-deficiency in the root environment and plant mineral nutrion. Pages 303-316.
In : Plant life under oxygen deprivation. M.R. Jackson, D. D. Davies, and H. Lambers, eds. SPB Academic
Publishing; The Hague, The Netherlands.
Drew M C, Chamel A, Garrec J P, Fourcy A (1980) Cortical air spaces (aerenchyma) in roots of corn
subjected to oxygen stress, structure and influence on uptake and translocation of 86Rb ions. Plant
Physiology. 65:506-511.
Drew M.C., Fourcy A(1986) Radial movement of cations across aerenchymatous roots of Zea mays
measured by electron probe X-ray microanalysis. J.Exp. Bot. 37: 823-831.
Drew M C, Guenther J, Lauchli A (1988) The combined effects of salinity and root anoxia on growth and
net Na+ and K+ accumulation in Zen mays L. (cv Pioneer 3906) grown in solution culture. Ann. Bot. 61:41-53.
Drew M C, Jackson M B, Giffard S (1979) Ethylene-promoted adventitious rooting and development of
cortical air spaces (aerenchyma) in roots may be adaptive responses to flooding in Zea mays L. Planta.
147:83-88.
Drew M C, Lauchli A(1985) Oxygen dependent exclusion of Na + by roots of Zea mays (cv. Pioneer 39060)
in relation to salinity damage. Plant Physiology. 79:171-176.
Drew M C, Lynch J M (1980) Soil anaerobiosis, micro-organisms and root function, Ann. Rev. Phytopath.
18:37-66.
Drew M C, Saglio P, Pradet A (1985) Higher adenylate energy charge and ATP/ADP ratios in
aerenchymatous roots of Zea mays in anaerobic media as a consequence of improved internal oxygen
transport. Planta 165:51-58.
Drew M C, Saker L R (1986) Ion transport to the xylem in aerenchymatous roots of Zea mays. J. Exp. Bot.
37:22-33.

77

Drew M C, Siworo E J (1979) The development of waterlogging damage in young barley plants in relation
to plant nutrient status and changes in soil properties. New Phytol. 82 :301-314.
Drew M C, Sisworo E J, Saker L R (1979) Alleviation of waterlogging damage to young barley plants by
application of nitrate and a synthetic cytokinin, and comparison between the effects of waterlogging,
nitrogen deficiency and root excision. New Phytol. 82: 315-329.
Ferguson I B, Clarkson D T (1976) Ion uptake in relation to the development of a root hypodermis. New
Phytol. 77:11-14.
Gambrell R P, Patrick W H (1978) The chemical and microbiological properties of anaerobic soils and
sediments. Pages 375-424. In : Plant life in anaerobic environments. D.D. Hook and R.M.M. Crawford,
eds. Ann.Arbor.Science, Ann. Arbor, Michigan.
Garcia-Novo F, Crawford R M M (1973) Soil aeration, nitrate reduction and flooding tolerance in higher
plants. New Phytol. 72 :1031-1039.
Greenway H, Waters I, Newsome J (1992) Effect of anoxia on uptake and loss of solutes in roots of wheat.
Aust. J. Plant Physiol. 19 :233-247.
Greenwood D J, Berry G (1962) Aerobic respiration in soil crumbs. Nature 195 :161-163.
Hanson J B (1978) Application of the chemi-osmotic hypothesis to ion transport across the root. Plant
Physiology. 62 :402-405.
Hodgson AS, Constable G R, Duddy Daniells I G (1990) A comparison of drip and furrow irrigated cotton
on a cracking clay soil 2. Water use efficiency, waterlogging, root distribution and soil structure. Irrig. Sci.
11:143-148.
Jackson M B, Fenning T M, Drew M C, Saker L R (1985a) Stimulation of ethylene production and gas-space
(aerenchyma) formation in adventitious roots of Zea mays L. by small partial pressures of oxygen. Planta.
165:486-492
Jackson M B, Fenning T M, Jenkins W (1985b) Aerenchyma (gas- space) formation in adventitous roots
of rice ( Oryza sativa L.) is not controlled by ethylene or small partial pressures of oxygen. J. Exp. Bot.
36:1566-1572.
Jacoby B, Rudich B (1980) Proton-chloride symport in barley roots. Ann. Bot. 46 :493-498.
Johnson J, Cobb B G, Drew M.C. (1989) Hypoxic induction of anoxia tolerance in roots of Zea mays. Plant
Physiology. 91 :837-841.
Leyshon A J, Sheard R W (1974) Influence of short-term flooding on the growth and plant nutrient
composition of barley. Can. J. Soil Sci. 54 :463-473.
Meyer W S, Barrs H D, Smith R C G, White N S, Heritage A D, Short D L (1985) Effect of irrigation on
soil oxygen status and root and shoot growth of wheat in a clay soil. Aust. J. Agri Res. 36 :171-185.
Peterson C A(1987) The exodermal casparian band of onion roots blocks the apoplastic movement of
sulphate ions. J. Exp. Bot. 38 :2068-2081.
Peterson C A, Perumalla C J (1984) Development of the hypodermal casparian band in corn and onion
roots. J. Exp. Bot. 35 :51-57.
Ponnamperuma, F N (1972) The chemistry of submerged soils. Adv. Agron. 24 :29-96.
Ponnamperuma, F N(1984) Effects of flooding on soils. Pages 9-45. In : Flooding and plant growth. T.T.
Kozlowski, ed. Academic Press, Orlando, Florida.
Poole R J (1978) Energy coupling for membrane transport. Annu. Rev. Plant Physiology. 29 :437-460.
Pradet A, Bomsel J L (1978) Energy metabolism in plants under hypoxia and anoxia. Pages 89-118. In :
Plant Life in anaerobic environments. D.O. Hook and R.M.M. Crawford, eds. Ann. Arbor, Michigan.

78

Rao K P, Rains D W (1976) Nitrate absorption by barley: 1. Kinetics and energetics. Plant Physiology.
57:55-58.
Reggiani R, Brambila U, Bertani A (1985) Effect of exogenous nitrate on anaerobic metabolism in excised rice
roots: 1. Nitrate reduction and pyridine nucleotides pools. J. Exp. Bot. 36 :1193-1199.
Robards A W, Clarkson D T, Sanderson J (1979) Structure and permeability of the epidermal/hypodermal
layers of the sand sedge (Carex arenario L.) Planta 101 :331-347.
Roberts J K M, Lane A N, Clark R A, Nieman R H (1985) Relationships between the rate of synthesis of
ATP and the concentrations of reactants and products of ATP hydrolysis in maize root tips, determined
by 31P nuclear magnetic resonance. Arch. Biochem. Biophys. 240 :7l2-722.
Roberts, J K M, Callis J, Jardetsky O, Walbot V, Freeling N(1984) Mechanism of cytoplasmic pH regulation
in hypoxic maize root tips and its role in survial under hypoxia. Proc. Nat. Acad. Sci. USA 81:3379-3383.
Saglio P H, Drew M C, Pradet A(1988) Metabolic acclimation to anoxia induced by low (2-4 kPa partial
pressure) oxygen pretreatment (hypoxia) in root tips of Zea mays. Plant Physiology. 86 :61-66.
Saglio P H, Raymond P, Pradet A (1983) Oxygen transport and root respiration of maize seedlings. A
quantitative approach using the correlation between ATP/ADP and the respiration rate controlled by
oxygen tension. Plant Physiology 72 :1035-1039.
Sanders D (1990) Kinetic modeling of plant and fungal membrane transport systems. Annu. Rev. Plant
Physiology. 42 :77-107.
Serrano R(1989) Structure and function of plasma membrane ATPase. Annu. Rev. Plant Physiology. 40 :61-94.
Shone M G T, Clarkson D T, Sanderson J (1969) The absorption and translocation of sodium by maize
seedlings. Planta. 86 :301-314.
Thomson C J, Greenway H(1991) Metabolic evidence for stelar anoxia in maize root exposed to low O2
concentrations. Plant Physiology. 96 :1394-1301.
Thomson C J, Armstrong W, Waters L Greenway H(1990) Aerenchyma formation and associated oxygen
movement in seminal and nodal roots of wheat. Plant Cell Environ. 13: 395-403.
Thomson C J, Greenway H(1991) Metabolic evidence tor stelar anoxia in maize roots exposed to low O2
concentrations. Plant Physiology 96 :1294-1301.
Trought M C T, Drew M C (1980a) The development of waterlogging damage in wheat seedlings (Triticum
aestivum L.) L. Shoot and root growth in relation to changes in the concentration of dissolved gases and
solutes in the soil solution. Plant and Soil. 54 :77-94.
Trought M C T, Drew M C (1980b) The development of waterlogging damage in wheat seedlings (Triticum
aestivum L.).II. Accumulation and redistribution of nutrients by the shoot. Plant and Soil 56 :187-199.
Trought M C T, Drew M C (1980c) The development of waterlogging damage in young wheat plants in
anaerobic solution cultures. J. Exp. Bot. 31 : 1573-1585.
Trought MC T, Drew MC (1981) Alleviation of injury to young wheat plants in anaerobicsolution cultures
in relation to the supply of nitrate and other inorganic nutrients. J. Exp. Bot. 32:509-522
Van Hoorn J W (1958) Results of a ground water level experimental field with arable crops on clay soil.
Neth. J. Agric. Sci. 6:l-10.
Van't Woudt B D, Hagan R M (1957) Crop responses to excessively high soil moisture levels. Pages 514-611.
In : Drainage of agricultural lands, J. N. Luthin. ed, : American Society of Agricultural Engineering,
Madison, Wisconsin.
Yeo A R, Yeo M E, Flowers T J(1988) Selection of lines with high and low sodium transport from within
varieties of an inbreeding species; rice (Oryza sativa L.). New Phytol. 110 :13-19.

79

RESPONSE OF FLOODING IN GOGORANCAH RICE

AND MOISTURE STRESS EFFECT AT REPRODUCTIVE
STAGE IN WALIK JERAMI RICE
Didiek Setio Budi and
Bambang Suprihatno1
SUMMARY
The effect of time of flooding and reproductive -stage moisture stress on crop
productivity was studied at the Jakenan experimental station during 1989-90 wet
season, and 1991 and 1992 dry seasons. The soil is classified as Planosol with a sandy
loam texture. Three separate field experiments were conducted in split plot design
with four replications. Time of flooding treatments consisted of flooding at, a) 4 wk,
b) 5 wk, c) 6wk, d) 7wk, e) 8 wk and f)9 wk after seeding. The reproductive stage
moisture stress treatments consisted of, a) continuous field saturation, b) noirrigation
from panicle initiation (PI) to heading and c) no irrigation from heading to harvest.
The rice genotypes used were Dodokan and Cikapundung for the gogorancah
experiment and IR64, IR36, Ciliwung, IR39357 - 71-1-1-2-2, IR 39422-18-1-2-2,
IR45912-9-1-2-2, IR48563-11-2-2-3, S969b-265-1-4-1, S400b-55-2 for the walik jerami
experiment (dry season).
The results showed that flooding at 7 week after seeding was very important
for gogorancah rice in achieving higher yields. Without rainfall during walik jerami
growing period, the moisture stress imposed from panicle initiation to heading
reduced grain yield up to 50%. When the moisture stress was imposed from heading
to maturity, yield reduction was around 70%. The results of 1991-92 experiment
failed to detect any of the effects of reproductive stage moisture stress on grain yield
because the rainfall was well distributed during the treatment period. In greenhouse
experiment the -50 cb moisture stress from panicle initiation to heading reduced
grain yield of IR64 by 50%. However, Ciliwung variety was relatively stable under
various stages of moisture stress. IR 39357-71-1-1-2-2 showed the intermediate
tolerance to the reproductive stage moisture stress and was also suitable for
gogorancah crop, particularly when the rainfall was uneven during the early part
of the growing season.
INTRODUCTION
In Pati district, rainfed rice area occupies about two-third of the total wetland
rice areas (irrigated rice 20,598 ha; rainfed rice 39,383 ha), and comprises 11% of the
total rainfed rice concentrated in Jaken, Jakenan, Winong and Puncakwangi
sub-districts.
1 SURIF, Jakenan Experiment Station, Jakenan, Pati. Central Java, Indonesia

The traditional cropping pattern at Jaken sub-district is gogorancah (dry

seeded bunded)during wet season, followed by walik jerami rice (minimum tillage,
transplanted) in dry season I and dryland crops in dry season II.
The average yield of gogorancah is around 6.0 t/ha under normal conditions.
However, variation in the onset and lack of rainfall before and after sowing are
serious production constraints for gogorancah rice. Drought often occurs in the
beginning of wet season, therefore, the farmers delay tillage and sowing. Drought
stress during reproductive stage poses a major problem for the walik jerami rice also
and so the yield often falls as low as 2.0 t/ha. The proportion of the length of dry
and wet periods during gogorancah growing season determines growth and yield
performance.
Little information is available on the time of flooding for gogorancah, although
many studies on the effect of moisture stress at different growth stages on rice have
been conducted. However, the results varied with location, season and variety. These
variations are attributed to the effect of climatic, soil and hydrological factors on
yield components and root growth and development.
The present study was conducted to (1) know the effect of time of flooding
on the yield of gogorancah rice in order to determine the most crucial time of flooding
for it and (2) to understand the reaction of varieties or breeding lines to water stress
at reproductive stage in walik jerami rice.

MATERIALS AND METHODS

Experiment 1. Effect of time of flooding on the yield of gogorancah rice.
A field experiment was conducted during 1988-84 wet season at the Jakenan
Experimental Station on the planosol soil with sandy loam texture, pH 5.6, organic
matter 0.6296, total N 0.07%, CEC 5.3 meq per 100 g soil, extractable P 18.0 mg Kg-1
and exchangeable K 0.04 meq per 100 g. soil,
A split plot design with four replications was used. Time of flooding was
studied on the sub-plot and variety on the main-plot. A very early maturing rice
variety Dodokan and an early variety Cikapundung were used. Time of flooding
consisted of flooding at, a) 4wk, b) 5wk, c) 6wk, d)7wk and e) 8wk and f) 9 wk after
seeding. Watering was from the rainfall through a water collector located near the
experimental site.
Rice seeds were dibbled at @ three seeds per hole, with 20 x 20 cm spacing.
The plot size was 4.0 x 5.0 m. All plots were fertilized at the rates of 112 kg N, 45
kg P 2O5 and 40 kg K2O ha-1. Nitro en fertilizer (Urea) was applied in splits, 40kg
N ha-1 at seeding time, 40kg N ha-1 3 wk after seeding and 32 kg N ha-1 at panicle
initiation (PI). Total amount of Potassium (KCI) and Phosphorus (triple
superphosphate; TSP) were given at seeding time together with the first nitrogen
application. Weeds, pest, and diseases were controlled intensively.
Observations were made for root dry mass, yield components and grain yield.
Root mass was sampled by a core method for determining the root mass density.
81

Yield Components and yield were recorded at harvest. Yield was obtained from 5
m2 harvest area and then adjusted to ton per hectare at 14% moisture content. Five
plants per plot were used to characterize the yield components.
Soil matric potential was measured by tensiometers' installed at 10 cm depth,
whereas perched water table was measured with piezometer inserted at the 75 cm
depth. Daily rainfall data were taken from a weather station located 50 m away from
the experimental site.
Experiment 2. Study on the effect of soil-water stress at reproductive stage on
grain yield in walik jerami rice.
Field experiments were conducted at the Jakenan Experimental Station during
1991 and 1992 dry seasons. A split plot design with three replications was used for
both experiments. In 1991, five rice genotypes IR 393557-71-1-1-2-2, IR39422-181-22,
lR459129-1-22,lR48563-11-2-2-3 and Dodokan were used as subplot and three levels of
soil water stress as the main-plot factor. The soil moisture stress treatments were : a)
continuous field saturation, b) no irrigation from panicle initiation(PI) to heading, and c)
no irrigation from heading to maturity. In 1992 rice, the genotypes used were:IR36,
IR39357-71-1-1-2-2, S969b-265-1-4-1 and S400b-55-2.
The 21-d-old seedings were transplanted at 2 seedlings/hill with 20 x20 cm
spacing. Land was ploughed twice by hoeing and harrowing. The plot size was 4 x
5 m and each plot received Urea), TSP and KC1 with the rates 135 kg N, 45 kg P2O5
and 50 kg K2 O ha-1 , respectively. Nitrogen fertilizer (Urea) was applied in three
equal splits of 45 KgN ha-1 each, at transplanting/ at 3 weeks after transplanting,
and at panicle initiation. All of KC1 and TSP were applied as basal. Weeds, pests
and diseases were controlled intensively.
Observations were made for yield components and yield. Four plants per plot
were used to characterize yield components. Grain yield was obtained from 8 m2
harvest area and calculated as ton per hectare at 14% seed moisture content.
Soil moisture status at 0-15 cm depth was determined by gravimetric method
and expressed as per cent oven dry weight (%ODW).
Experiment 3. Greenhouse study on the effect of soil water stress at different
growth stages on the yield of lowland rice.
This experiment was conducted at the Sukamandi Experimental Station during
1989 dry season using Ultisol soil with silty clay texture. A split plot design with four
replications was used. Four levels of soil water stress were used as sub-plot factor i.e.:
a) continuous flooding, b) water stress imposed from 10 DAT to panicle initiation, c)
water stress imposed from panicle initiation to heading and d) water stress imposed
from heading to maturity. Two varieties were used as main plot factor i.e. a) IR64 and
b) Ciliwung. The soil moisture stress was allowed to reach -50 cb and measured by
tensiometer at 10 cm depth. The water stress was imposed by not watering the pots.
After reaching the stress level, pots were reirrigated until field saturation. The 21-d old
seedlings were transplanted to each pot with 4 hills per pot.
82

Grain yield data were obtained from each pot and expressed in gram per hill
at 14% moisture content.
Crop data were analyzed using the methods for the split plot design and Duncans
Multiple Range Test was employed if F-test showed significance.
RESULTS AND DISCUSSION
The results of the effect of time of flooding on the yield components and grain
yield for gogorancah rice are presented in Table 1. Panicle length, 1,000-grain weight
and grain yield were affected by time of flooding. Flooding at four to six weeks after
seeding was important period to achieve higher yields. There was a tendency of
decline in grain yield when the flooding was applied more than six weeks after
seeding. Spikelet number per panicle and filled-spikelet percentage were not affected
by time of flooding treatments. There were significant differences in spikelet number
per panicle, filled- spikelet percentage and 1,000- grain weight but not in grain yield
between Dodokan and Cikapundung cultivars.
The decline in yield at, or above seven weeks after seeding might be due to
low supply of water and nutrients particularly nitrogen to the crop. Ismunadji (1985)
stated that the form and availability of nutrients are directly related to moisture
supply in the soil.
The effect of time of flooding on the mot mass density for the gogorancah rice
is illustrated in Figures 1 and 2.

Fig. 1 :

Root growth of Dodokan variety at the 0-5 cm soil depth as affected by time of flooding,
Jakenan, 1988-89 WS.

83

Fig. 2 :

Root growth of Cikapundung variety at the 0-5 cm soil depth as affected by time of flooding,
Jakenan, 1988-1989 WS.

Table 1.

Effect* of time of flooding and cultivars on the yield components and

grain yield for gogorancah rice, Jakenan, 1988-89 WS.

Treatments

Cultivars (C)
Dodokan
Ci kapund ung
Time of flooding (F)
Weeks after seeding
(WAS)
4 WAS
5 WAS
6 WAS
7 WAS
8 WAS
9 WAS
CV (C):
CV (F):
*

Yield components
Spikelet/
panicle (nos)

Filled
spikelet
(%)

1,000
grain wt
(g)

Panicle
length
(cm)

84.3 b
119.4 a

75.0 a
69.5 b

24.7 a
23.5 b

22.3 b
23.1 a

105.4 a
98.3 a
103.0 a
105.1 a
101.5 a
97.6 a

74.4 a
73.5 a
70.7 a
71.9 a
71.1 a
72.0 a

5.3%
7.8%

4.7%
4.5%,

25.1 a
24.5 ab
24.2 abc
23.8 bc
23.3 c
23.6 a
3.3%
3.8%

23.4.a
22.9 ab
22.9 ab
22.4.b
22.3 b
22.3 b
3.9%
3.3%

In a column, means followed by a common letter are not significantly different at

the 5% level by DMRT

Grain
yield
(t/ha)

4.17 a
3.95 a

4.54
4.24
4.10
4.05
3.81
3.63

a
ab
ab
bc
bc
c

18.9%
10.0%

There was gradual increase in root mass density from 37 to 88 days after
seeding. The rooting characteristics were influenced by water regime and type of
cultivars. Longer dry period relative to wet period tended to increase the root mass
density at 0-5 cm depth.
Apparently Cikapundung had higher root mass density at 0-5 cm depth than
Dodokan, particularly when flooding was done at nine weeks after seeding. Rooting
depth of both the varieties was deepest when grown under dryland condition.
An aerobic soil favours deep root growth and rice varieties differ markedly in
their rooting habits, both laterally and vertically (Yoshida and Hasegawa, 1982). The
fluctuations in soil matric potential and piezometric water level during gogorancah
growing season are illustrated in Figures 3 and 4.

Fig. 3 :

Fluctuation of the soil matric potential at the 10 cm soil depth and piezometric water level
for the two time of flooding of Dodokan. Jakenan, 1988-1989 WS.

High fluctuation in soil matric potential occurred from 25 to 40 days after seeding
and was subsequently relatively stable. The piezometic water level was close to the soil
surface from 45 days after seeding till harvest.
The dynamics of soil matric potential and the depth of piezometric water level
are partly controlled by rainfall. Weekly rainfall well distributed was from 45 days
afterseeding until harvest (Fig.5) and therefore, the soil matric potential and perched
water table were found stable.
85

Fig. 4 :

Fluctuation of the soil matric potential at the 10 cm soil depth and piezometric water level
for the two time of flooding of Chillwung, Jakenan, 1988-1989 WS.

The effect of rice genotypes and soil water stress at the reproductive stage on
the yield components and yield are presented in Tables 2 to 5. The results of genotype
x water stress interaction on the number of spikelets per panicle are presented in
Table 2. Dodokan and IR3922-18-1-2-2 were found to have stability in number of
spikelets per panicle under reproductive moisture stress.
Table 2.

Interactive effect between rice genotypes and soil water stress at

reproductive stage on the number of spikelets of rice, Jakenan, 1991 DS.

Treatment
Rice genotypes (C)

Reproductive stage soil water stress (S)

Continuous
saturation
(no stress)

Primordia
to heading

Heading to
maturity

Spikelet/panicle (no)*

IR393537-71-1-1-2-2
IR39422-18-1-2-2
IR45912-9-1-2-2
IR48563-11-2-2-3
Dodokan
CV (C):
CV (S):

86

96.7 a
69.4 b
51.7 c
56.3 c
80.5 b

74.4 b
99.9 a
60.7 c
68.4 bc
89.9 a

78.1 a
82.7 a
62.0 b
64.0 b
75.9 a

12.07%
9.3%

In a column, means followed by a common letter are not significantly different at the 5% level by
DMRT

The number of spikelets per panicle were shown to be reduced by reproductivephase drought (De Datta, 1981). The effects of moisture stress at the reproductive stage on
the yield components and yield are presented in Table 3.
Table 3.

Yield components and grain yield as affected by rice genotypes and soil
water stress at reproductive stage, Jakenan, 1991 DS.

Treatment

Soil water stress (S)

Continuous saturation

Yield Components*
Filled
spikelets (%)

1,000 grain
wt (g)

Grain yield
(t/ha)

65.8 a

1.96 a
0.97 b
0.36 c

PI to heading
Heading to maturity
Rice genotypes (C)

47.8 b
24.7 c

26.6 a
26.0 a
23.1 b

IR393537-71-1-1-2-2

51.3 a

25.9 a

1.20 a
1.19 a

IR39422-18-1-2-2

52.0 a

22.8 b

IR45912-9-1-2-2

40.8 b

25.6 a

0.99 b

IR48563-11-2-2-3

44.4 b

25.6 a

0.97 b

Dodokan

42.1 b

26.3 a

1.13 ab

CV (S) :
CV (C) :

16.0%
13.4%

5.3%

5.0%

In a column, means followed by a common letter are not significantly different at the 5% level by
DMRT.

During the growing season of walik jerami rice, occurrence of moisture stress
from panicle initiation to heading and from heading to maturity, significantly influenced
the filled spikelet percentage and grain yield. Moisture stress from panicle initiation to
heading and from heading to maturity reduced yield by 50% and 72%, respectively.
However, moisture stress from heading to maturity significantly influenced the 1,000
grain weight.
Moisture stress in the early reproductive stage (panicle initiation to heading)
resulted in a decreased grain yield through Education in number of spikelets.
Moisture stress during reproductive phase has been shown to reduce filled spikelet
number per panicle (IRRI, 1990). However, it had no influence on the panicle number
per hill, 1,000 grain weight and yield of walik jerami rice during the dry season of
1992 (data not shown).

87

Fig. 5 :

Weekly rainfall distribution in gogorancah rice growing season, Jakenan, 1988-1989.

The genotype IR36 exhibited greater spikelet sterility under both the
continuous saturation and reproductive moisture stress situation. Apparently, soil
water stress from heading to maturity induced the spikelet sterility (Table 4).
Table 4.

Interactive effect of rice genotypes and soil water stress at reproductive

stage on the spikelet sterility of rice, Jakenan, 1992 DS.

Rice genotypes (C)

Soil water stress (S)

Continuous
saturation

Panicle initiation
to heading

Heading to
maturity

Spikelet Sterility (%)*

41.0 a
34.3 a

IR36

15.8 a

IR393537-71-1-1-22

13.3 a

23.9 ab

28.8 b

S969b-265-1-4-1

10.2 a

16.2 c

17.2 b

S400b-55-2

18.7 a

23.0 ab

19.5 b

CV (C) :

29.6%

CV (S) :

31.0%

88

In a column, means followed by a common letter are not significantly different at

the 5% level by DMRT

The rice genotypes grown in 1992 walik jerami did not show any significant
differences in grain yield. Water stress treatments at reproductive stage had no effect
either. The average yield ranged from 2.2-2.6 t/ha. The genotypes IR36, however,
exhibited a tendency to yield higher (though not significant) than other breeding
lines (data not presented). The rainfall distribution was optimum during the
stress-treatment periods and hence the soil moisture status reached above field
capacity. This is considered as critical level of soil moisture for rainfed lowland rice
(Fagi, 1986), below which significant yield differences could be expected.
A comparison between two genotypes (grown in 1989) to moisture stress
tolerance revealed that the level of yield of Ciliwung was lower than IR64.
However, it was found to be more tolerant to moisture stress at various growth
stages. IR64 was more sensitive to moisture stress from panicle initiation to
heading than other phases of moisture stress, as evidenced by a drastic reduction
in grain yield (Table 5).
Table 5.

Interactive effect of rice cultivars and soil water stress at different

growth stages on the grain yield (g/ hill), Sukamandi Greenhouse, 1989
DS.

Soil water stress (S)

Cultivar (C)
IR64

Ciliwung
grain yield* (g/hill)

Continuous

flooding

20.82 a

17.26 a

10 DAT to Panicle initiation

18.82 a

15.71 a

PI to Heading

10.31 b

15.04 a

Heading to maturity

19.90 a

17.25 a

CV (C) :

3.0%

CV (S) :

10.7%

In a column, means followed by a common letter are not significantly different at the 5% level by
DMRT

CONCLUSION
In Pati district, the optimum time of flooding in gogorancah rice is in the
range of four to six weeks after seeding to obtain higher yield. The moisture stress
from panicle initiation to heading was found to be critical in walik jerami rice. The
genotypes Ciliwung and TR39357-71-1-1-2-2 were found, having potential to be
planted as Walik jerami rice.
89

REFERENCES
De Datta S K (1981) Principles and practices of rice production. John Wiley & Sons.
Fagi A M (1986) Efficient use of water for rainfed lowland rice. In : Progress in rainfed lowland rice. IRRI,
Los Banos, Philippines.
IRRI (1990) Program report for 1989. International Rice Research Institute., Los Banos, Philippines.
Ismunadji M (1985) Effect of sulphate application on the performanceof IR36 rice variety under submerged
and dryland condition. Indonesian Journal of Crop Science l(1) : 1-64.
Yoshida S, S Hasegawa (1982) The rice root system, its development and function. In : Drought resistance
in crops with emphasis on rice. IRRI, Los Banos, Philippines.

90

ANAEROBIC TOLERANT RICE SEEDLING GROWTH

AND ITS POTENTIAL IN THE TROPICAL RAINFED
LOWLAND ECOSYSTEMS
M. Yamauchi 1 , A.M. Aguilar 1 , J.L Padilla 1
D.V Aragones1 and U. Singh2

SUMMARY
Rice seedling establishment in wet direct seeding is erratic because farmers
sow seeds on the soil surface. Anaerobic- tolerant cultivars that grow vigorously
in flooded or water- saturated soil were identified. These cultivars developed
longer coleoptiles than non-tolerant controls, suggesting that they can transport
02 in the air or water to the seeds in anaerobiosis efficiently. The superiority of
tolerant cultivars in seedling survival and growth persisted even when grown in
the dark, although their growth increased under light. Temperatures between 15
and 35C had little effect on difference in seedling survival between tolerant and
control cultivars, although seedling growth of both types of cultivars was reduced
at 15C.
The growth of a tolerant cultivar in flooded soil was equivalent to that of a
control grown in drained soil. Thus, the use of anaerobic cultivars may make sowing
in flooded or water-saturated soil similar to that in freely drained soil. The potential
of its applicattion to rainfed lowland ecosystems was tested in Tarlac, Philippines,
during 1993 wet season. Rice seeds soaked in water for 24 h were dibble-sown into
water-saturated soil under zero tillage. There was little difference between
anaerobic-tolerant and control cultivars in crop establishment, presumably due to
soil dryness (there was no rainfall after seeding). The tolerant cultivars outyielded
the controls: the six anaerobic cultivars had a mean grain yield of 3.2 t/ha at 0-180
kg N/ha, whereas the control cultivars had 2.4 t/ha. Further agronomic and
physiological
studies
are
needed
to
adapt
anaerobic-tolerant cultivars
in
heterogeneous rainfed lowland soils.

INTRODUCTION
Crop establishment
A rice crop is established either by transplanting or by dry or wet direct
seeding. Transplanting is mostly practiced in irrigated lowland rice and is an
established technology. Dry seeding is also well studied and is practiced where

1 International Rice Research Institute, P.O. Box 933, 1099 Manila, Philippines

2 International Fertilizer Development Center-IRRI cooperative Program

soil is dry. Rice plants adapted to irrigated lowland and upland ecosystems are bred
for transplanting and dry seeding, respectively, in Asia.
Wet seeding technology, however, has not been well developed. Germinated
seeds are sown on the surface of drained, water- saturated puddled soil or flooded
soil surface (water seeding). Crop establishment by wet seeding seems to be more
unstable and inconsistent than that by dry seeding. In dry seeding, seeds are covered
with aerobic soil and protected from birds, rodents, and rain splashing; in wet
seeding, seeds are sown on puddled soil surface and exposed to these disturbances.
The condition of the puddled soil surface changes according to intensity of puddling,
climate, and time after land preparation. In water seeding, seedlings not only drift
due to poor anchorage but are also destroyed by snails and fish.
Rice seedlings grown by wet seeding are exposed to soil anaerobiosis, although
this is not a constraint in transplanted and dry seeded rice. Because there are no
cultivars raised and adapted for wet seeding, farmers currently use cultivars
developed specifically for transplanting.
Anaerobic seeding is proposed to improve the wet seeding process (Yamauchi
et al., 1993a). Germinated seeds are sown under the surface of puddled soil with or
without standing water so that seeds are protected from the disturbances occurring
on the soil surface. In addition, seeds get better anchorage in water seeding. Because
seeds are exposed to soil anoxia, cultivars tolerant of anoxia are preferable for this
seeding method.
The process of crop establishment in wet seeding could be divided into survival
(which can be measured by percentage seedling establishment) and growth (which can
be characterized by leaf development, seedling height, and shoot and root weights).
Mesocotyl elongation is an important factor in raising the apical meristem to the upper
soil layer when seeds are sown deep into the soil (Takahashi, 1984).

Germplasm selection
A screening system to find rice germplasm suitable for anaerobic seeding was
developed (Yamauchi et al., 1993a). Seeds that germinated for 2 d were planted 25
mm deep in seedling trays. The trays were then submerged in water to a depth of
30-50 mm. Seedling establishment was evaluated by analyzing leaf development,
seedling height, and percentage establishment (survival) 15 d after planting.
Rice germplasm tolerant to soil anoxia were identified in traditional cultivars and
also in breeding lines with improved plant type. Screening for anaerobic-tolerant rice
germplasm was conducted using 256 accessions from conserved germplasm from the
International Rice Germplasm Center (IRGC) and 404 accessions from the International
Network for Genetic Evaluation of Rice (INGER), IRRI (Yamauchi et al., 1993 a). IRGC
germplasm represented broad genetic diversity, and INGER germplasm involved
desirable agronomic characters. Eight percent of the IRGC germplasm and 2% of those
from INGER were identified statistically as tolerant of soil anoxia. Among the anaerobic
germplasm were those from northeast India and Bangladesh which were adapted to
deepwater and early summer rainfed lowland cultures.
92

In vitro growth characteristics

Germinated seeds were placed in vials or erlenmeyer flasks with distilled
water and exposed to various degrees of hypoxia. The anaerobic-tolerant cultivars
had longer coleoptiles than the control cultivars in the gas flow of N2 and air, and
underhypoxia in closed flasks in the dark (Yamauchi and Herradura, 1993; Yamauchi
et al., 1994). Although ethylene stimulated coleoptile elongation of tolerant and
control cultivars was at the same degree, mesocotyl and shoot (leaf and sheath)
elongations of tolerant cultivars were more stimulated by ethylene compared with
those of control cultivars. The tolerant cultivars produced more ethylene per plant
than did the control cultivars; at times, however the same amount of ethylene were
produced. Although the in vitro seedling growth of tolerant cultivars was more than
that of control cultivars, the fundamental mechanism underlying the vigorous
growth is unknown.
Growth characteristics in soil
Seedling establishment may be partly controlled by some factors related to
soil properties and soil management. In the Philippines, crop establishment was
better in upland soil than in lowland soils, presumably due to higher redox potential
in the upland soil when the soils were flooded (Yamauchi et al., 1993b).
The presence of chloroplasts in the coleoptile was demonstrated by Kordan
and Ashraf (1990), suggesting that photosynthetically produced O2 in the coleoptile
may be involved in the process of seedling survival and that assimilated
carbohydrates may accelerate seedling growth. Because the coleoptile is the first
organ to reach the surface of the flooded soil, it may be important to confirm if
photosynthesis is involved in the process of seedling establishment.
Because temperature controls the catabolism of seed reserves, it may have an
effect on seedling survival and growth. The growth of anaerobic-tolerant cultivars
in aerobic soil has not been compared with the growth in anaerobic soil.
Characterizing anaerobic seedling growth as a function of temperature and water
may help to determine the rice ecosystem in which anaembic-tolerant cultivars may
be adopted.
Anaerobic seeding in irrigated lowlands
Rice seeds sown under the surface of puddled soil can successfully establish
seedlings particularly when anaerobic- tolerant cultivars are used. Broadcast seeds
can be placed under the soil surface when sowing is done immediately after
puddling. Anaerobic drill seeding is practiced by using a seeder. Anaerobic broadcast
seeding and drill seeding were successfully conducted in the Philippines during dry
and wet seasons in two locations (Aragones et al., 1993, 1994). The most serious
constraint identified in this study was the destruction of seedlings by freshwater
snails ( Pomacea sp.) (Mochida, 1988). Anaerobic broadcast seeding was also
successfully conducted in the Mekong Delta (Chau and Yamauchi, 1994) and in
Hanoi (Chuong and Yamauchi, 1994) in Vietnam. Anaerobic seeding in irrigated
lowland is feasible.
93

The yield potential of anaerobic tolerant cultivars with improved plant type
was equivalent to or higher than that of high-yielding local check cultivars when
grown under direct seeding. In yield trials at Los Baos, Philippines, the
anaerobic-tolerant cultivar IR52341-60-1-2-1 had the same high yield as check IR72
(8.4 t/ha). At Muoz, IR52341-60-1-2-1 yielded more than IR 72 (9.8 vs 7.0 t/ha). In
the Mekong Delta, anaerobic-tolerant cultivar BR1870-67-1-3 yielded 5.8 t/ha,
whereas check MTL103 yielded 5.1 t/ha during the 1993 wet season (Chau and
Yamauchi, 1994). IR52341-60-1-2-1 yielded 5.2 t/ha in Hanoi with check CN2 yielding
4.5 t/ha during the 1993 wet season (Chuong and Yamauchi, 1994).
Wet seeding in the rainfed lowland ecosystem
Because soil water conditions in rainfed lowland fields vary according to
location in toposequence, rainfall, and time after the onset of the wet season, farmers
choose dry seeding, wet seeding, or transplanting in order to establish the crop
successfully and economically. Drought is a serious constraint in rainfed lowland
rice culture. Early planting by direct seeding advances the plant developmental stage,
thus avoiding drought at the termination of rainfall at the end of the wet season.
Direct seeded plants have both seminal and nodal roots, whereas transplanted plants
have only nodal roots. In addition, seminal roots might develop deeper than nodal
mots at the early growth stage. Thus, direct seeded rice might be more tolerant of
drought at the early stage than transplanted rice.
Compared with dry seeding, wet seeding is less commonly practiced in the
rainfed lowland fields, presumably due to the instability of crop establishment.
Success in wet seeding depends on the characteristics of the dry-to-wet transition
period at the onset of the monsoon and on adequate surface drainage (Morris and
Zandstra, 1979). Water seeding is practiced where drainage is difficult. Introducing
anaerobic seeding in the rainfed lowlands would stabilize crop establishment in wet
seeding, and thus may interest farmers to follow direct seeding.
Seedling growth of anaerobic-tolerant cultivars in flooded soil is characterized
here in terms of light and temperature, and in comparison with growth in drained
soil in controlled environments. We also studied the crop establishment and grain
yield of rice plants in rainfed lowland fields with water- saturated soil under zero
tillage to identify the contraints in adoption of anaerobic seeding.
MATERIALS AND METHODS
Anaerobic seedling growth under continuous light and darkness
Seeds of three anaerobic-tolerant cultivars (CO25 [percentage germination,
93.4%; rate of germination, 1.00], ASD1 [99.6; 1.001, and Taothabi [98.3; 1.001 and
three control cultivars IR36 [96.5; 1.00 IR50 [95.2; 0.951, and IR74 [91.7; 0.99]) were
germinated in petri dishes for 2 d at 30C, and then sown in sieved (mesh 40) upland
(Maahas clay) soil placed in a container (300 x 240 x 110 mm). Rate of germination
was measured according to Krishnasamy and Seshu (1989). The container was filled
with soil up to a height of 80 mm. Seventeen seeds were sown for each cultivar per
94

container. The soil was then submerged under 30 mm of water. Eight containers
were placed in a growth chamber (30C, 70% relative humidity) under continuous
darkness or light (30 k lux). Plants in four containers (replications) were sampled 7
and 11 d after seeding.
Seedling growth was characterized by measuring emergence score (0=no
emergence, 1=coleoptile emerged, 2=lst leaf emerged, 3= 2nd leaf emerged, ...) ;
seedling height; and mesocotyl length of the individual plant; and shoot, mot, and
seed weights as a bulk sample per replication after drying at 80C. Percentage of
seedlings established was calculated to estimate seedling survival.
Effect of temperature on anaerobic seedling growth
Seeds of six cultivars (CO25 [percentage germination, 93.7%; rate of
germination, 0.991, ASDl [l00; 1.001, Taothabi [97; 0.941, IR 36 [95.5; 0.981, IR50 (93.3;
0.931, and IR74 [91.3; 0.991) were germinated and planted using the method described
above. The containers were placed in temperature-controlled glass rooms (outdooor
growth chamber) (under natural light) using the set of 15, 20 and 30C or 25, 30, and
35C. One container represented one replication. Four containers (replications) were
placed in each glass room. Plants were sampled 14 d after seeding and their growth
and survival analyzed.
Seedling growth in flooded and drained soils
Seeds of IR36 (percentage germination, 96.8%; rate of germination, 0.98) and
Taothabi (98.9; 0.97) were similarly germinated and sown, except that the plastic
containers were replaced by 3.8-liter pots. The soil in the pot was flooded at a depth
of 5 cm or freely drained, Seventeen seeds of the cultivar were sown per pot. The
pots were placed in a temperature- controlled (29/2lC day/night) glass room
(phytotron) under natural light . Plants were sampled 4, 6, 8, 10, 12, and 14 d after
seeding and their growth was measured. Root length (from the tip of seminal root
to the base of plant) was measured. The experimental design was a randomized
complete block with four replications (pots).
Performance in rainfed lowlands
Seeds of six anaerobic-tolerant cultivars (IR341996-50-2-1-3 [percentage
germination, 99.3; rate of germination, 1.00], IR52341-60-1-2-1 [100; 1.00],
IR50363-61-1-2-2 [99.3; 0.991, IR 31802-48-2-2-2 [100; 1.00], BR1870-67-1-3 [98.5; 0.89],
and BR736- 20-3-1 [100; 1.00]) and two controls (PSBRC10 and IR72) were
dibble-sown under zero tillage in a rainfed lowland field at Masalasa, Tarlac,
Philippines, during 1993 wet season. All seeds were produced during 1993 dry
season; dormancy was broken by keeping seeds at 50C for 5 d.
After dry-season harvest, the remaining straw was removed from the field.
The seeds were soaked in water for 1 d, then dibbled manually at 5-10 seeds per
hole at a depth of 25-30 mm at 20x 20cm spacing. Soil was saturated with water at
sowing and dried after 1 wk. Weeds were removed by hand. Fertilizer was applied
at the rate of 0, 60, 120, and 180 kg N/ha in four splits (15 and 35 d after emergence,
95

5 d before panicle initiation, and flowering). Experimental design was a randomized

complete block with four replications. Plot size was 6 x 4 m.
Crop establishment was measured 25 d after seeding by counting the number
of seedlings in a 0.8 m2 area per plot. Seedling dry weight, mesocotyl length, and
emergence scores were taken from the plants, using 4 hills/plot. Grain yield was
measured from a 5m2 area.

RESULTS AND DISCUSSION

Anaerobic seedling growth under continuous light and darkness
This study was conducted to clarify how light affects anaerobic seedling
growth and if light reaction is involved in the mechanism of cultivar difference in
the tolerance for soil anoxia. Seedling growth was enhanced in the light presumably
due to photosynthesis (Fig. 1).
Emergence score, seedling height, and shoot and root weights were greater
under light than under darkness. On the other hand, seedling survival was not
affected by light. Means of percentage seedling establishment of control and
anaerobic cultivars were 67.2 and 94.4% under light and 71.6 and 94.9% under
darkness, respectively. This indicates that light reaction is not involved in the process
of seedling survival.
The percentage seedling establishment at 7 d of IR36 under light and of IR50
and IR74 under darkness decreased at 11 d; those of the others were not changed.
This suggests that seedling establishment had been determined before 7 d and that
some seedlings established at 7 d died later. Lack of photosynthesis is not the cause
of seedling death.
Regardless of light conditions, there were significant differences between
control and anaerobic-tolerant cultivars in percentage seedling establishment,
emergence score, seedling height, shoot and root weights, and mesocotyl length
(Fig.1). The results suggest that light is not involved in the mechanism of cultivar
difference.
Effect of temperature on anaerobic seedling growth
This study was conducted to determine the effect of temperature
survival and growth. The percentage of seedling establishment was less
temperature than was seedling growth, which was measured by shoot
height (Fig.2). Growth significantly diminished at 15C but was highest

on seedling
affected by
weight and
at 30C.

The difference in seedling growth between the control and anaerobic-tolerant

cultivars diminished at 15C, whereas the difference in survival was little affected
by the change in temperature. These results suggest that the factors controlling
seedling growth may be different from those controlling seedling survival.
Crop establishment in the field is determined by biomass production per unit
land area, which is the product of the number of seedlings per unit area (determined
96

Fig. 1.

Rice seedling growth in flooded soil under continuous light and darkness at 30C. Seedling
growth was measured at 7 and 11d after seeding. Bars having a common letter (under light
or dark condition at 7 or 11d) are not significantly different at the 5% level by DMRT.
Number denotes cultivar names: 1=IR36; 2=IR50; 3=IR74; 4=CO25; 5=ASD1; 6=Taothabi.
The shaded and closed bars indicate the light and dark conditions, respectively.

97

by seed rate and seedling survival) and single-seedling weight. Therefore, crop
establishment may be low at lower temperatures and the highest at 30C.

Fig. 2.

98

Effect of temperature on rice seedling growth under natural light condition. Growth was
measured 14d after seeding. Experiments were conducted 2 times with the temperature
combinations of 15, 20, and 30C (closed symbols) and 25, 30 and 35C (open symbols).
Values are the means of control (IR36, IR50. and IR74) ( ) and anaerobic tolerant cultivars
(CO25, ASD1, and Taothabi) (O). The difference in each character between control and
tolerant at each temperature was significantly different at the 1% level (even at 15C for
seedling height and shoot dry weight).

Seedling growth in flooded and drained soils

This study was conducted to compare growth of the control and
anaerobic-tolerant cultivars in drained and flooded soils. Taothabi was superior to
IR36 in terms of seedling height, shoot weight, and mesocotyl length. Emergence of
IR36 was inhibited in the flooded soil, whereas Taothabi was not (Fig.3), an indication
of its anaerobic character.

Fig. 3.

Seedling growth and seed weight of rice cultivar IR36 (control) ( ) and Taothabi (anaerobic)
(O) in flooded (closed) and drained soils (open). Values having a common letter at each
day are not significantly different at the 5% level by DMRT.

99

Development in shoot weight and seedling height suggested that, although

growth in drained soil was superior to that in flooded soil, the growth of Taothabi
in flooded soil was equivalent to that of IR36 in drained soil. The root length of
Taothabi in flooded soil was inhibited until 6 d after seeding, but the roots started
to grow later. After 8 d, root length of Taothabi was the same as that of IR36 in the
drained soil. It could be presumed that in Taothabi, which was grown in flooded
soil, the O2 in the atmosphere was transported from the shoot to the root between
the 6th and 8th d and then was utilized to develop the roots (since emergence score
was more than 2 (1st leaf emerged) at 6 d).
The decrease in seed weight, which may indicate the catabolism of the seed
reserved material for use in energy production in the tissue, was larger in drained
soil than in flooded soil. The larger decrease in seed weight in Taothabi than in IR
36 in drained or flooded soil suggests rapid metabolic turnover in Taothabi.
Performance in rainfed lowland field
This field experiment was conducted to analyze yield potential of
anaerobic-tolerant cultivars and to identify the constraints in growing them under
the rainfed lowlands. After harvesting the dry-season rice crop, the rain started and
soil was saturated with water. The land was too wet to conduct dry seeding, although
it did not have enough standing water to puddle the soil for transplanting. Under
these field conditions, farmers generally waited for more rainfall and thus planting
gets delayed. Delayed planting increased the chance for plants to encounter drought
because rain often stops at the late growth stage. Introduction of zero-tillage wet
seeding may advance the time of planting over transplanting. We assumed that
sowing anaerobic-tolerant cultivars using zero tillage may be possible because they
can survive and establish seedlings even with heavy rains and even if fields were
flooded after sowing.
There was no heavy rainfall aftersowing. Soil dried up and a crust was formed.
There was no significant difference between anaerobic-tolerant and control cultivars
in terms of plant density (Table 1). Similarly, no significant differences in emergence
score, seedling height, shoot weight per unit area, and mesocotyl length were
observed.
Grain yield of anaerobic-tolerant cultivars was higher than controls
irrespective of the dose of N application (Table 1). IR4 1996-50-2-1-3 produced the
highest yield (3.0 t/ha) among the cultivars tested at 0 kg N/ha while BR736-20-3-1
showed highest yield (4.6 t/ha) at 180 kg N/ha. The percentage increase in grain
yield by the introduction of anaerobic tolerant cultivars was 79% at 0 kg N/ha and
23-30% at 60-180 kg N/ha.
IR72 is a high-yielding cultivar in irrigated lowlands. Hence, the low yield of
IR72 (1.5-2.9 t/ha) demonstrates that the yield potential was hindered by
unidentified stresses and that the anaerobic-tolerant cultivars were more tolerant to
such stresses than IR72. Therefort, it could be presumed that anaerobic-tolerant
cultivars have the tolerance for the unidentified stresses. The screening system
should be able to select rice germplasm tolerant not only of soil anoxia but also of
100

unidentified stresses. The results suggest that stress tolerance in rainfed lowland rice
is important in increasing the yield.
Table 1.

Crop establishment* and grain yield* of 6 anaerobic tolerant and 2

control cultivars in rainfed lowlands, Masalasa, Tarlac, Philippines. 1993
wet season.

Cultivars

Crop establishment**
Height
(mm)

Grain yield (t/ha)

Plant
density

Shoot
dry wt.

Mesocotyl
length
(mm)
0

(no./m 2 )

(g.m 2 )

IR4199650-21-3

124bcd

10.5

216

5.4

IR5234160-1-2-1

147ab

12.9

169

IR5036360-1-2-2

116cd

8.1

IR31802
48-2-2-2

143abc

BR187067-1-3

N levels (kg/ha)
60

120

180

3.0a

3.3ab

3.7a

3.3bc

5.4

2.4abc

2.9abc

3.3ab

4.0ab

170

5.0

2.1bcd

2.9abc

3.5a

3.4bc

12.3

181

7.3

2.0cd

3.5a

3.6a

3.8ab

96d

7.2

149

5.1

2.8ab

3.0abc

3.7a

3.3bc

BR73620-3-1

102d

9.3

191

7.1

2.0bcd

2.9abc

3.5a

4.6a

PSBRC10
(Control)

169a

12.6

177

6.9

1.1e

2.6bc

3.1ab

2.9c

IR72
(Control)

123bcd

12.5

176

6.6

1.5de

2.3c

2.7b

2.9c

In a column, means followed by a common letter are not significantly different at the 5%
level by DMRT.

**

Measured 25 d after seeding.

RESEARCH NEEDS
The results presented here demonstrated that anaerobic-tolerant cultivars had
superiority in seedling establishment over a wide range of environmental conditions
: light, temperature, drained/flooded soil. In addition, they were more tolerant to
the unidentified stresses in the rainfed lowland field in the Philippines, producing
more grains than the non-tolerant controls. Introducing anaerobic cultivars into
rainfed lowlands would increase not only the stability of wet seeding but also the
grain yield. Systematic research is needed to develop wet seeding technology in the
rainfed lowlands, in as much as crop establishment and grain yield are influenced

101

by many factors-- climate, soil, seed vigor, and cultivar:

*
*
*
*
*

The mechanism of the superiority in seedling survival and growth and

high grain yield of anaerobic-tolerant cultivars need to be clarified.
The data base of available anaerobic-tolerant germplasm for use in
agronomic and breeding studies should be developed.
Breeding of high-yielding anaerobic-tolerant cultivars is necessary.
Genetic studies including RFLP analysis will be useful.
There is a need to search for germplasm with multiple stress tolerance
because of exposure of germinating seeds to anoxia and drought in the
rainfed lowland ecosystem.
There is a need to develop proper agronomic practices of crop
establishment for the transition time between dry seeding and
transplanting to maximize use of the potential of cultivar traits.
Zero-tillage wet seeding technology should be studied to enable early
planting of rice.
REFERENCES

Aragones D V, Casayuran P R, Asis C A, Aguilar A M, Sta Cruz P C, Yamauchi M (1993) Evaluation of

crop stand establishment by anaerobic seeding technology at two locations in the Philippines. Philip. J.
Crop Sci. 18: 36 (suppl. 1).
Aragones D V, Casayuran P R, Cruz R T, Yamauchi M (1994) Direct seeding technology: drill seeding
using anaerobic cultivars. Philip. J. Crop Sci. 19: 21 (suppl. 1).
Chau N M, Yamauchi M (1994) Performance of anaerobically direct seeded rice plants in the Meakong
Delta of Vietnam. Int. Rice Res. Notes 19: 6.
Chuong P V, Yamauchi M (1994) Anaerobic direct seeding of rice in Northern Vietnam. Paper presented
at the Wet Seeded Rice Workshop, 31 May-3 June 1994, Bangkok, Thailand.
Kordan H A, Ashraf M (1990) Environmental anoxia is unnecessary for inhibiting chloroplast
photomorphogenesis in rice coleoptile ( Oryza sativa L.). J. Exp Bot. 41: 435-440.
Krishnasamy V, Seshu D V (1989) Seed germination rate and associated characters in rice. Crop Sci. 29:
904-908.
Mochida O (1988) Non-seed borne rice pests of quarantine importance. Pages 117-129 In :Rice seed health.
Internattional Rice Research Institute, P.O. Box 933, Manila, Philippines.
Morris R A, Zandstra H G (1979) Land and climate in relation to cropping patterns. Pages 255-274 In :
Rainfed lowland rice. International Rice Research Institute, P.O. Box 933, Manila, Philippines.
Takahashi N (1984) Seed germination and seedling growth. Pages 71-88 In : Biology of rice. S. Tsunoda
and N. Takahashi, eds. Japan Science Society press, Tokyo.
Yamauchi M, Aguilar A M, Vaughan D A, Seshu D V (1993a) Rice ( Oryza sativa L.) germplasm suitable
for direct sowing under flooded soil surface. Euphytica 67: 177-184.
Yamauchi M, Aguilar A M, Sta Cruz P S (1993b) Anaerobic seeding with suitable germplasm. Int. Rice
Res. Notes. 18: 36.
Yamauchi M, Herradura P S (1993) Superior coleoptile elongation of rice varieties suitable for anaerobic
seeding. Int. Rice Res. Notes. 18: 37-38.
Yamauchi M, Herradura P S, Aguilar A M (1094) Genotype difference in rice postgermination growth
under hypoxia. Plant Sci. Lett. (in press)

102

CARBOHYDRATE STATUS OF RAINFED LOWLAND

RICES IN RELATION TO SUBMERGENCE, DROUGHT
AND SHADE TOLERANCE
G.S. Chatuverdi1, P.C. Ram1 , A.K. Singh1 , P.Ram1
K T. Ingram3 , B.B. Singh1 , R.K. Singh1 and V.P. Singh2

SUMMARY
Rice accumulates significant amount of carbohydrate in different plant parts
especially in the culm which has an important role when plant experiences drought
and flooding at various stages of crop growth. Carbohydrate status of rainfed
lowland rice vis-a-vis mechanism of submergence tolerance at seedling stage,
drought at flowering stage and shade at reproductive stage has been discussed with
experimental evidences. Tolerance to submergence at seedling stage, and drought
at flowering was closely associated with higher levels of carbohydrates (total water
soluble carbohydrate and starch). Hydrolytic enzyme, alpha amylase activity had
positive correlation with seedling survival at initial stage of submergence and was
negatively correlated during submergence. Yield of rice under water deficit at
booting or flowering was directly related to carbohydrate content, apparent
translocation rate (ATR) and apparent contribution rate (ACR). Chemical desiccation
method for screening flowering stage drought resistance has been discussed.
Histochemical analysis technique for screening for flowering stage drought tolerance
in rice has also been discussed. Adaptive features of wild rice species Oryza
australiensis in relation to drought has been discussed which include reduced tiller
and leaf area, more carbohydrate in culm and rhizomes.
INTRODUCTION
Rainfed lowland rice represents one of the major rice ecosystems of South
and South-East Asia. About one-fourth of the world rice area falls under this category
and contributes 19% of total rice grain yield. In India, approximately 14 million
hectare rice lands are rainfed lowland, of which major area (60%) comes from Eastern
India. Eastern Uttar Pradesh constitutes about 66% of rainfed lowlands in the state
(Table 1).
The productivity of rainfed lowland rice is low and averages around 1.0-1.5
t/ha which, though is higher than upland and deep water rice, but much lower than
that of irrigated ecosystem. Productivity of rainfed lowland is about 48% of value
product per hectare of irrigated rice (ADB, 1989). Among the general causes of poor
productivity of rainfed lowland are (i) poor establishment of seedlings due to
1

N.D.University of Agriculture and Technology, Kumarganj, Faizabad, U.P., India.

International Rice Research Institute, P.O. Box 933, Manila, Philippines
3 Peanut CRSP Management Office, University of Georgia, Griffin, Georgia, USA.
2

partial to complete submergence (Chaturvedi et al., 1993; Vergara, 1985), (ii) late
season drought/anthesis drought (Chaturvedi and Ingram, 1989, (iii) low light stress
in wet (Kharif) season (Murty and Sahu, 1987; Venkateswarlu and Srinivasan, 1978),
(iv) accumulation of toxic substances in ill-drained soils (Ponamperuma, 1984), (v)
soil factors such as salinity/alkalinity and zinc deficiency (Rao and Biswas, 1979),
(vi) cold stress at panicle initiation (Nanda and Coffman, 1979) and (vii) severe
occurrence of diseases and pests etc. Among the causes mentioned above,
submergence and drought are the most prevalent stresses to reduce yield, and
account for about 70% and 25% of yield loss, respectively in rainfed lowland rice
area (Widawsky and O'Toole, 1990). Thus evaluating and screening cultivars for
submergence at initial stage and for drought at later stages of crop growth become
an important pre-requisite to improve overall productivity of rainfed lowland rice.
Besides, there is a need to screen for salt, shade and cold tolerance for
location-specificity.
Table 1 .
Region

Total rice and rainfed lowland rice area by regions

Total Rice Area

Rainfed Lowland

(000 HA)

(000 HA)

Rice Area

World

146400

37950

Asia

131000

36200

India

41000

14601

Eastern India

24000

8880

Uttar Pradesh

5500

2245

Eastern U.P.

2661

1415

Limited studies have been done on the morphological, physiological,

biochemical and anatomical bases of submergence tolerance (Levitt, 1980; Singh,
1982; HilleRisLambers and Seshu, 1982. Vergara, 1985; Setter et al., 1989 a & b;
Chaturvedi et al., 1993), flowering stage drought resistance, and shade tolerance
in rice (Murty and Sahu, 1987; Venkateswarlu, 1977; Chaturvedi and Ingram,
1989).
In the present paper, the aim is to review and discuss one of the physiological
factors i.e. carbohydrate status of rainfed lowland rice vis-a-vis mechanism of
submergence tolerance at seedling stage and drought at flowering stage.
Experimental results of rainfed lowland rice research consortium, conducted at
Masodha Faizabad, India and of flowering stage drought stress behaviour, observed
at IRRI, Manila, Philippines are discussed.
104

A REVIEW
Carbohydrate status and stress tolerance
Most of the carbohydrates are produced by green leaves and other green
tissues of plants and are either translocated as sugar or stored temporarily as sugar,
starch or fructans. Later in the dark, sugar is resynthesized and translocated from
the leaf. In most plants, sucrose is translocated via phloem to carbohydrate sinks for
cell growth, metabolism, respiration or storage. Growing cells act as carbohydrate
sinks but later they may act as source when they become photosynthetically active.
At subcellular level chloroplast may act as carbohydrate source and mitochondria
of same cell act as sink (Kozlowski, 1992).
Growth and grain yield depend on both current photosynthesis and stored
assimilates. If current photosynthesis is limited by environmental stresses such as
water deficit (Marshall et al., 1989 Bunce, 1982: Blum et al., 1983; Chaturvedi and
Ingram, 1988), low temperature (Hilliard and West, 1970; Chatterton et al., 1972),
nutrient deficiency (Lenhert et al., 1979), or submergence (Setter et al., 1989 a and b;
Chaturvedi et al., 1993), the remobilization of stored assimilates is accelerated and
plant is forced to depend on stored assimilates. Stored assimilates also enhance
recovery of plants after stress (Wardlaw and Eckhardt, 1987).
Stored carbohydrates have important role in metabolism, growth,
development or stress tolerance, defense and survival (Kozlowski, 1992). It is
therefore, also called reserve carbohydrate (RC) and is particularly sensitive to late
season stress and management practices. It is also essential for survival when
photosynthesis is low or is stopped. Among stored carbohydrates, starch is the most
important reserve carbohydrate. Starch accumulates whenever high level of sugars
build up and starch is transformed into sugars when sugars are low (Kozlowski and
Keller, 1966). Among soluble carbohydrates, sucrose is the major translocable and
storage carbohydrate. Reserve carbohydrates are also important in preventing
invasion by certain plant pathogens and insect attacks.
Carbohydrate status and submergence tolerance
Rice crop suffers from partial to complete submergence in rainfed lowland
areas. Complete submergence occurs during short- term flash flooding, varying upto
15 days, near river valleys and riverlets. Characteristics of flood water vary at
different locations and the nature depends upon light intensity under water, turbidity
and concentration of dissolved gases such as O2, CO2 and C2H4. Partial submergence
of rice is quite common in low lying areas.
Complete submergence of rice leads to low concentration of carbohydrates
and reduced growth and finally death of tissues (Palada and Vergara, 1972). It is
partly due to low rate of photosynthesis (Trought and Drew, 1980; Rai and Murty,
1979), which is directly associated with slow diffusion of CO2 in water (Setter et al.,
1989 a). Other factors which contribute to slow growth are (i) acceleration of
photorespiration, (ii) anaerobiosis leading to breakdown of CHO and carbon loss as
CO2 and ethanol, (iii) high ethylene production disturbing membrane integrity and
105

chlorosis, (iv) leakage of CHO, amino acids and ketones, (v) ABA accumulation and
decreased supply of cytokinins which lead to stomatal closure and (vi) reduction of
RUBP carboxylase activity. Complete submergence reduces level of carbohydrates
but in partial submergence the reduction depends on proportion of leaves below
the water (Setter et al., 1989). Submergence reduces concentration of soluble sugars
and starch in all plant parts of rice by 4-12 fold (Setter et al., 1989).
Amount of carbohydrate in plant parts correlates positively with submergence
tolerance (Palada and Vergara, 1972; Emes et al., 1987; Chaturvedi et al., 1993). Older
seedlings which have greater submergence tolerance, also contain more carbohydrate
(Vergara, 1985; Chaturvedi et al., 1993). Decreasing seedbed soil nitrogen level
increases seedling carbohydrates and submergence tolerance as well (Vergara, 1985).
Emes et al. (1987) observed that under submerged condition, the culm of
submergence-tolerant varieties accumulated higher levels of starch and at a more
rapid rate than did varieties with low submergence tolerance. Culms of submergence
tolerant plants contained starch even after being submerged for 7 days, whereas the
starch reserve of submergence susceptible varieties was exhausted during the same
period. Plant may use reserve starch stored in the culm during submergence, and
the ability to survive for longer periods may be related to continued availability of
carbohydrates. Chaturvedi et al. (1993) showed that levels of total water soluble
carbohydrates and starch were positively correlated with tolerance to
complete/partial submergence. Tolerant genotypes had lower rate of reduction of
CHO during submergence and higher recovery after 30 days. Slow rate of reduction
in tolerant genotypes makes the substrate available for regeneration of growth after
releasing the submergence. Alpha- amylase activity was also higher before
submergence and after recovery in tolerant genotypes. In contrast, under partial
submergence condition, leaf sheath plays the major role in carbohydrate distribution.
High carbohydrate depletion rate (CDR) in leaf sheath was observed in tolerant than
susceptible genotypes under partial submergence condition.
Submergence for 3 days under continuous light led to mobilization of 65% of
the starch from the regions of rice internodes which has been formed prior to
submergence and disappearance of starch was accompanied by a 70-fold
enhancement of amylolytic activity (Raskin and Kende, 1984). Submergence also
caused 26-fold increase in the translocation of newly synthesized photosynthate
assimilate from leaves to internodes and younger regions of culms. Tops of rice
contain significantly larger amounts of starch than those of submerged plant
(Yamaguchi, 1973). Amylase activity increases in rice leaves and internodes during
submergence (Yamaguchi and Sato, 1963).
Carbohydrates status and drought tolerance at reproductive stage
Substantial quantities of carbohydrates are accumulated in different parts in
rice (Perez et al., 1971) which are of paramount importance when plant experiences
water deficit at flowering (Rahman and Yoshida, 1985; Chaturvedi and Ingram, 1988,
1989). Pre-anthesis assimilates stored in stem and leaf sheath tissue appear to
contribute substantially to grain filling under water stress (Akita, 1987; Cock and
106

Yoshida, 1972; Chaturvedi and Ingram, 1989). In rice, translocation of assimilates

increases under stress (Reyniers et al., 1982). Pre-anthesis assimilates in wheat are
also known to contribute to grain filling when drought occurs in the later stages of
growth (Gallagher et al., 1976).
Water stress causes reduction in the pool size of sucrose in many crop species
(Lawlor and Fock, 1978). This reduction is not due to decrease in the level of enzymes
but due to overall changes in the direction of metabolic fluxes. More organic carbon
is sequestered for the production of amino acids such as glycine, serine and
glutamine.
Water stress reduces net photosynthate availability by reducing leaf area and
increasing stomatal resistance. This is followed by a decrease in activity of enzymes
such as RUBP Carboxylase and in the photochemical activity of chloroplast (Sinha
et al., 1982; Marshall et al., 1980).
Maintenance of functional integrity and survival under stress depends upon
the capacity of plant for osmoregulation (Turner and Jones, 1980; Hanson and Hitz,
1982). This is commonly associated with starch breakdown and concomitant increase
in low molecular weight solutes. Major contributing solutes to osmoregulation
comprise soluble sugars, free amino acids, organic acids, betains and inorganic ions
and polyols (Morgan, 1984). Genetic differences exist in degree of osmotic adjustment
in sorghum genotypes (Ackerson et al., 1980).
Translocation of photosynthates within the plant is more tolerant to water
stress than growth and photosynthesis (Boyer, 1976). The relative importance of
translocation in supplying carbohydrates to grain increases when stress develops
(Hunt, 1979; Austin et al., 1980). Varieties differ for stored stem carbohydrates and
their rate of translocation into the grain increases under stress. Upto 31% of rice
grain yield under stress was accounted for by the translocation of stem reserves.
Differential ability of genotypes to sustain translocation-based grain growth under
stress by destroying the photosynthetic source at the grain filling stage was assesed
in wheat (Blum et al., 1993 a, b; Regan et al., 1993). This was done by spraying field
plots with chemical dessicants (magnesium chlorate, atrazine and KI solution) that
kill leaf blades, awns and glumes. Injury by late season water stress was correlated
with injury by chemical dessication.
Carbohydrate status and shade tolerance
Reduced light is one of the major constraints to the productivity of rainfed
lowland rice due to prevalent clouds in monsoon season, as it is mostly grown during
wet season (July to December). Low light during primordial initiation (PI) to
flowering, or flowering to harvest, reduces grain development considerably
(Venkateswarlu, 1977; Kato, 1986), while low light at anthesis leads to high spikelet
sterility and low harvest index. Under low light at flowering, carbohydrate decreases
and GA and soluble N in panicle increase, leading to a high spikelet sterility. Greater
accumulation of carbohydrate in shoots at flowering relates to tolerance to low light
stress (Murty and Sahu, 1987). Under low light, stored carbohydrates contribute
substantially to grain filling (Janardhan et al., 1980). Under shade, total sugars (mostly
107

non reducing sugars and starch) are markedly reduced in all plant parts. Total soluble
N and amino N are enhanced in susceptible varieties. Higher sterility is ascribed to
disturbed N metabolism and accumulation of soluble N in the panicle which is toxic
to normal grain setting (Murty et al., 1983).

EXPERIMENTALRESULTS
Submergence tolerance
Submergence tolerance in rice is not governed by a single factor. Our previous
results (Chaturvedi et al., 1993) indicated that elongation and submergence tolerance
are two independent traits. Attempt has been made to combine these two traits in
one genotype (Raysaha et al., 1993). Submergence tolerance is a dominant character
and additive gene effects appear to be operative in the control of this trait (Mohanty
and Khush, 1985).
From our three years (1991-1993) data, it is clear that initial high carbohydrate
level in the tissues is one of the adaptive traits of submergence tolerance. Higher
levels of initial carbohydrate (water soluble sugar and insoluble starch) act as buffer
stock and its continued availability during submergence is critical for the survival
and growth of rice. Metabolic energy needed by the plant during submergence is
primarily supplied from stored carbohydrates present in the tissue.
Tolerant genotypes had higher total initial carbohydrate content and starch
before submergene and had lower rate of depletion during submergence than the
susceptible (Fig. 1 & 2). Susceptible genotypes had poor accumulation of CHO and
showed faster rate of depletion during submergence. Slow rate of depletion of CHO
in tolerant genotypes seems to be one of the adaptive traits to withstand submergene
due to continuous availability of substrate for energy for a longer period. In
susceptible plants, drastic reduction of CHO leads to high rate of anaerobic
fermentation and production of ethanol at toxic level. Setter et al. (1989b) also
proposed the accumulation of end products to toxic levels in their model on response
of rice to submergence.
Carbohydrate status of plant changes with the type and nature of
submergence, plant age and exposure to water phase (Setter et al., 1989b). In one of
the experiments, plants were completely submerged and both shoots and roots were
under water, carbohydrate depletion occurred and it was more in susceptible than
in the tolerant genotypes.
Tolerant genotypes had higher recovery of total carbohydrate and starch 30
days after releasing submergence stress than the susceptible ones. The alpha amylase
activity was higher before as well as after releasing stress (Fig. 3). Higher alpha
amylase activity in leaf and culm seemed important for longer submergence duration
tolerance in tolerant genotypes. In general, older seedlings (45- day old) had better
survival than younger seedling (15, 30-day old) and it was due to higher level of
stored CHO in older seedlings. Vergara (1985) also observed higher level of CHO
before submergence which decreased at low rate during submergence.

108

Fig. 1.

Changes in starch content of tolerant and susceptible genotypes in complete submergence

(15 days). Pooled data of 1992-93.

109

Fig. 2.

110

Changes in total water soluble sugar (WSS) in tolerant and susceptible genotypes in complete
submergence (15 days). Pooled data of 1992-93.

Fig. 3.

Changes in a amylase activity in tolerant and susceptible genotypes in complete submergence

(15 days). Pooled data of 1992-93.

111

Contrarily, in deep water rice, Emes et al., (1987) reported that culms of
submergence tolerant varieties maintained higher levels of starch, and its accumulation
was more rapid than the less tolerant type. Raskin and Kende (1984) working with deep
water rice variety-Habibganj Aman from Bangladesh, found that starch disappearence
was acompanied by 7-fold enhancement of amylolytic activity. Internodes of air grown
rice plants contained large amounts of starch (about 10% DW) which disappeared upon
submergence in water under continuous light. This indicates that carbohydrate status
and its remobilization alter with the status of plant in water.
Correlation studies showed that survival of rice was positively correlated with
the level of total water soluble sugar and starch in the culm, before submergence,
after submergence and after recovery (Table 2). Strong positive correlation was
found in shoot (culm) after submergence stage rather than before and after recovery.
There was a negative correlation between alpha amylase and survival in both mot
and shoot after submergence.
Table 2 .

Character

Correlation values between survival and starch, survival and total water
soluble sugar and survival and alpha amylase activity in different plant
parts. 30 d old seedling, 11 d submergence. Pooled data of 1992-1993.
Relationship

r value*
B.S.

A.S.

A.R.

N.S.
0.95**

N.S.
0.99**

0.99**
0.96**

Survival and total water soluble sugar

Root
Shoot

N.S.
O.80**

N.S.
0.76**

N.S.
0.92**

Survival and alpha-amylase activity

Root
Shoot

0.85
0.94

-0.90
-0.99

N.S.
N.S.

Survival and starch

Root
Shoot

BS=Before Submergence ; AS=After Submergence ; AR=After Recovery

Drought tolerance
Rice accumulates significant amount of carbohydrates before heading and
there exists wide variability among genotypes. Improved rices have higher contents
of total non-structural carbohydrate at heading than do the low yielding traditional
varieties (Weng et al., 1986). Previous experiments have shown that when water
deficit occurs during flowering or grain filling, grain growth derives a greater
112

proportion of substrates from stored assimilates than when no stress occurs.

Assimilates accumulated prior to anthesis are of paramount importance when plant
experiences water deficit at later growth stages.
In general, cultivars which had higher levels of CHO before flowering, also
had higher yields under both normal and water deficit conditions. A strong positive
correlation was observed between flowering stage CHO in all plant parts and yield
(Table 3). A non-significant relationship was found between tissue CHO at tillering
and yield. Carbohydrate contents of different plant parts were positively correlated
with filled grains/hill at flowering stage (Table 3). Leaf blade carbohydrate had
greater affinity for increasing filled grain, as evidenced by high r value (0.78).
Table 3 :

Correlation coefficient between filled grain/hill and carbohydrate

content and yield and carbohydrate content of different plant parts
(stress and control data combined).
Time of water deficit

Plant Parts

Tillering

Booting

Flowering

Filled grain/hill versus carbohydrate content

Leaf blade

0.23

0.42*

0.78**

Leaf sheath

0.20

0.49*

0.51*

Culm

0.14

0.19

0.46*

Yield versus carbohydrate content

Leaf blade

0.21

0.47

0.85**

Leaf Sheath

0.20

0.41*

0.66**

Culm

0.20

0.46*

0.52**

and **

significant at p 0.05 and 0.01, respectively.

Yield was inversely Correlated with apparent translocation rate (ATR). Among
cultivars, yield was directly correlated with maximum ATR under stress condition,
and cultivar IR-46 maintained better translocation under stress than did the other
cultivars (Table 4).
One of the goals of plant breeders is to produce germplasm with stable yields,
and a measure of yield stability is the slope of ATR vs yield regression, the smaller
the slope, the more stable the yield. Variety IR-64 had the greatest slope and high
crop susceptibility (cs) factor.

113

Table

4.

Apparent translocation rate (ATR) and

apparent
(ACR) in lowland rices in response to water deficit.

contribution

rate

Cultivars

Time of
water deficit

ATR

ACR

Yield
g/plot

Crop
sucesptibility

IR64

Control
Tillering
Booting
Flowering

0.10
0.17
0.24
0.29

0.13
0.21
0.34
0.39

19.4
16.7
12.8
11.9

0.13
0.34
0.38

IR46

Control
Tillering
Booting
Flowering

0.14
0.21
0.31
0.42

0.16
0.26
0.44
0.51

22.3
19.2
15.7
14.7

0.06
0.22
0.23

Mahsuri

Control
Tillering
Booting
Flowering

0.08
0.15
0.19
0.25

0.11
0.16
0.26
0.32

15.9
13.2
11.1
09.8

Salumpkit

Control
Tillering
Booting
Flowering

0.08
0.13
0.22
0.31

0.11
0.16
0.31
0.42

14.3
11.8
10.1
11.1

0.17
0.29
0.22

IR54

Control
Tillering
Rooting
Flowering

0.12
0.19
0.27
0.36

0.14
0.22
0.37
0.40

20.3
18.3
13.8
11.8

0.13
0.25
0.31

0.04

0.05

0.6

0.05

LSD 0.05

0.17
0.31
0.39

In the histochemical studies also, it was observed that tolerant genotypes had
higher level of starch localization at flowering in non-stress environment than
susceptible genotypes. Histochemical observation of CHO localization was very
close to the observation recorded by quantitative method (Tables 5-7). Under stress,
carbohydrate is remobilized relatively faster in tolerant than in susceptible genotypes
as also evident by carbohydrate depletion rate (CDR) (Table 6). It is evident from
Table 7, that yield of tolerant cultivar IR 46 was relatively higher under stress
condition than that of susceptible variety IR 52. Higher level of sugar acts as osmotic
agent to maintain higher leaf area and dry weight. Poor yield in susceptible cultivar
IR 52 was mainly due to its poor storage of CHO and poor translocatory behaviour
which was reflected in its poor dry weight and, leaf area, and more unfilled spikelets
(64%) in comparison to tolerant ones (30%).

114

Table 5.

Culm carbohydrate content in three genotypes at flowering under

well-watered, (without stress), during stress, and after recovery.

Cultivars

Culm CHO mg/pot

Without stress

During stress (80 kPa)

After recovery

IR-46

3000

1390

580

N-22

2230

900

420

IR52

2300

1200

840

470

281

136

LSD 0.05

Table 6.

Culm carbohydrate depletion rate (CDR) in three

subjected to water deficit (80 kPa SMT) at flowering.

Cultivars

rice

genotypes

Culm CHO (mg/pot)

Control

Stress (80 kPa SMT)

IR-46

40

80

N-22

38

64

IR52

28

34

LSD 0.05

14

Table 7.

Yield and yield attributes of three rice genotypes under well-watered

(C) and water-deficit (WD) conditions (80 kPa) at flowering.

Cultivars

Grain yield
(g/pot)
C
WD

IR46

20.6

14.0

105

N-22

16.4

12.4

IR52

17.9

9.0

LSD 0.05

1.1

Spikelet/
panicle
WD

Individual
grain weight (mg)
C
WD

Unfilled
grain (%)
C
WD

83

23

20

16.5

30.5

87

71

19

18

20.0

30.5

107

87

24

19

19.7

3.0

1.1

63.8
3.1

Among wild species too, variation existed in carbohydrate accumulation and

remobilization (Table 8). Wild species Oryza nivara had higher accumulation of CHO
115

in leaf blade, while Oryza australiensis had higher accumulation of CHO in culm and
leaf sheath.
Table 8

Total carbohydrate content and carbohydrate depletion rate (CDR) of

wild rices under well watered and water deficit (80kPa) at flowering.

Wild rices

Water treatment

Leaf

CHO (mg/pot)
Culm

Oryza
australiensis

Control
Stress

1500
1050

3600
2340

1080
646

40
60

Oryza officinalis

Control
Stress

1800
1530

3000
2400

780
585

30
35

Oryza nivara

Control
Stress

2400
2160
430

2560
2048
570

750
600
375

40
48

LSD 0.05

Root

CDR
mg/pot/day

All wild species showed rhizomatous characters, full of starch, but Oryza
austruliensis had higher starch localization and remobilization than other wild species
as evident by histochemical and quantitative analysis. This was reflected in higher
yield under stress condition. Oryza nivara, which produced higher dry matter, did
not produce enough yield under stress. Among 3 wild species, Oryza australiensis
showed more adaptive traits like reduced tiller and leaf area, and more carbohydrate
content before stress and faster remobilization during stress. Higher intensity of
starch localization in rhizomes is a strong indicator of tolerance behaviour under
stress.
In one of the experiments, a chemical dessicant was used as an agent for
simulating drought to identify genotypes for drought tolerance at flowering.
Carbohydrate depletion rate in vegetative tissue gives a relative estimate of CHO
remobilization to panicles. Except for roots, IR 46 had the highest CDR in all plant
parts. Thus, drought tolerant IR 46 had both the traits of accumulation and
remobilization of CHO. Both traits together confer tolerance to water deficit or
chemical dessication at flowering. Cultivar N22 had the highest CDR in its roots.
High root and culm CDR may explain renewed tiller production at the expense of
grain production. Our results are in close agreement with the work on wheat done
by Blum et al, (1983) and in rice by Chaturvedi and Ingram, (1991). Submergence
tolerant cultivar had minimum CDR value, and therefore showed that faster
remobilization is not a feature of submergence tolerance.
All cultivars responded similarly to water deficit and chemical dessication
with respect to CHO distribution, leaf area, dry matter, and yield and yield
components. Injury index by drought was more closely related to injury index by
chemical dessication at 5 days after panicle emergence (DAPE) than at 10 days before
116

panicle emergence (DBPE). The results showed that chemical dessicant had an effect
on rice yield similar to that of water deficit, with the common effect, that of forcing
the plant to rely on previously accumulated assimilates for grain growth. Chaturvedi
and Ingram (1991) also found that low concentration of atrazine (0.05 kg a i/ha) in
a non-stress environment, was a potential simulator of drought.

Shade tolerance
In general, shade reduced CHO in all plant parts relative to unshaded control.
The greater reduction in CHO however, was observed when shaded at 5-20 DAPE
than 10-25 days after panicle initiation (DAPI) in both wet and dry seasons.
Reduction in CHO by shading was also reported by Eaton and Ergle (1954), and
Kemp (1981).
Table 9 .

Correlation coefficient of grain yield and carbohydrate content of

various plant parts at different stages of growth in lowland rice under
wet and dry treatments.

Stage

Treatment

Leaf blade

Culm

10 DAPI

WET

0.79**

0.72

0.53

DRY

0.75*

0.73

0.62

WET

0.45

0.34

0.59

DRY

0.35

0.36

0.62

WET

0.78**

0.76*

0.52

DRY

0.76*

0.80**

0.61

WET

0.48

0.29

0.09

DRY

0.22

0.25

0.02

WET

0.49

0.07

0.37

DRY

0.13

0.07

0.10

25 DAPI
5 DAPE
20 DAPE
HARVEST

Leaf sheath

* and ** Significant at P 0.05 and 0.01, respectively.

Grain yield was positively correlated with CHO in leaf blades and culm at 10
DAPI and 5 DAPE (Tables 9 and 10). The culm was the greatest source of translocated
CHO for grain growth is supported by ACR. The results indicate that under water
deficit, grain growth depended on previously accumulated assimilates. Grain yield
and ATR were negatively correlated, but correlation was positive with maximum
ATR value. Our results are similar to those of Reyniers et al. (1982) who also reported
higher ATR values under water deficit. Translocation in supplying CHO to grain
increases when plant experiences stress (Austin et al., 1980). Upto 31% of rice grain
yield under stress was accounted for by translocation of stem reserves. Among the
117

cultivars, IR 64, IR 46 and Mahsuri had similar ATR values with no shade but IR 46
had more CHO coupled with high ATR and ACR under both shade and water deficit
conditions. Thus, IR 46 had the best CHO accumulation and translocation to confer
stress tolerance.
Table 10.

Variety

Effect of water and shade on yield, apparent translocation rate (ATR),

and apparent contribution rate (ACR) in lowland rice.
Wet season

Treatments

Dry season

Yield

ATR

ACR

Yield

ATR

ACR

IR 46

No shade
Shade 10-25 DAPI
Shade 5-20 DAPE

4.0
3.1
2.8

5
10
24

20
21
35

3.9
3.1
2.8

13
22
34

20
23
40

IR 64

No shade
Shade 10-25 DAPI
Shade 5-20 DAPE

4.6
3.6
3.4

6
12
28

24
26
45

4.4
3.5
3.3

8
23
40

24
28
46

Mashuri

No shade
Shade 10-25 DAPI
Shade 5-20 DAPE

3.0
2.1
1.8

4
10
20

21
22
36

2.9
2.0
1.7

6
12
25

20
25
38

0.1

0.1

LSD 0.05

Reduction of rice grain yield by shade was also reported by Murty and Sahu,
1987; Venkateswarlu, 1977; Jadhav, 1987; Patro and Sahu, 1988). Yoshida (1973) found
that shade, particularly at reproductive and ripening phases reduced rice yields by
40 to 78 compared to unshaded controls. Additive effects of shade and water deficit
were observed in the dry but not in wet season trials. It was concluded that IR 46
had the greatest shade and drainage tolerance during panicle emergence and
flowering owing to its ability to form and set spikelets using accumulated assimilates
(Chaturvedi and Ingram, 1989).

CONCLUSION

118

Submergence tolerance at seedling stage is closely associated with higher

levels of water soluble carbohydrate and starch in the tissues. Survival of
seedlings was positively correlated with the levels of WSC and starch
before and after stress as well as after recovery.

Slow rate of depletion of CHO during submergence appears to be a

desirable trait for tolerance. This will help the plant to continue its growth
for a longer period of submergence.

Submergence tolerance and elongation ability are not related. Most

elongating genotypes are not submergence tolerant with exception to a
few genotypes which showed high submergence tolerance alongwith
moderate elongation ability.

Hydrolytic enzyme-alpha amylase-activity seems crucial in the

remobilization of assimilates, and had positive correlation with seedling
survival at initial stage and was negatively correlated during submergence
stage.

Culm is the major organ which accumulates higher stored carbohydrate.

Carbohydrate level in plant is significantly correlated with survival.

Higher CHO at anthesis plus high ATR appear useful traits to select
cultivars for drought tolerance at flowering.

Yield of rice under water deficit at booting or flowering was directly

related to the level of CHO, ATR and ACR.

Yield was directly correlated with maximum ATR under stress condition.

Drought tolerant IR 46 culm had higher intensity of starch staining before

stress and faster remobilization, during and after stress, than drought
susceptible cultivar IR 52.

Spray of a chemical dessicant, magnesium chlorate (4%), had similar

effects on CHO remobilization, yield components and yield as that of
water deficit.

Injury index by drought and chemical dessicant were significantly and

strongly correlated at 5 DAPE. Dessication method can be used to screen
large number of genotypes where water deficit is difficult to be imposed.

IR46 had relataively more shade and drainage tolerance during flowering
stage owing to its ability to form and set spikelets using accumulated
assimilates.

Histochemical analysis of starch localization and remobilization is quick

and reliable method and can be used as a tool of CHO qualitative analysis
in breeding programme.

Wild species Oryza australiensis, showed more drought adaptive traits

like reduced tiller, leaf area, more carbohydrate in culm and rhizomes
before stress and faster remobilization during stress.
REFERENCES

Ackerson R C, Krieg D R, Sung F J M (1980) Leaf conductance and osmoregulation of field grown sorghum
genotypes. Crop Sci. 20: 10-16.
ADB (1989) Rainfed Agriculture in Asia and pacific. Asian Development Bank Manila, Philippines. 644p.
Akita S (1987) Physiological aspects for improving yield potential in tropical rice cultivation. Paper
presented at International Rice Research Conference 21-25 September 1987, IRRI, Manila, Philippines.

119

Austin R B, Morgan R B, Ford M A, Blackwell R D(1980) Contribution to grain yield from pre-anthesis
assimilation in tall and dwarf genotypes in two contrasting seasons. Ann. Bot.. 45: 305-319.
Blum A, Poiarkova H, Golan A, Mayer J (1993) Chemical dessication of wheat plants as simulator of postanthesis stress. I.Effects on translocation and Kernel growth. Field Crops Res. 6: 149-155.
Boyer J S (1976) Post-anthesis at low water potential.Phils. Trans Royal Soc. Lond. Ser. B. 273 : 501-512.
Bunce J A (1982) Effect of water stress on photosynthesis in relation to diurnal accumulation of
carbohydrate in source leaves. Can. J. Bot. 60: 195-200.
Chaturvedi G S, Misra C H, Singh O N, Pandey C B, Yadav V P, Singh A K, Dwivedi J L, Singh B B, Singh
R K (1993) Physiological basis and screening for tolerance for flash flooding. Pages 78-96. In: Rainfed
lowland rice-agricultural research in high risk environments. KT Ingram, ed., International Rice Research
Institute, Los Banos, Philippines.
Chaturvedi G S, Ingram K T (1991) Effect of atrazine and moisture stress on translocation of assimilates
in rice. Plant Physiol. and Biochem. India 18 (1) 40-42
Chaturvedi G S, Ingram KT (1989) Growth and yield of lowland rice in response to shade and drainage.
Philip. J. Crop Sci. 14 (2) 61-67.
Chaturvedi G S, Ingram KT (1988) Assimilate translocation of lowland rice in response to water deficit
and shade. Paper presented at the 4th FCSSP annual meeting. 27-30 April, 1988, Davao city, Philippines.
Chatterton N J, Carlson G E, Hungerford W E, Lee D R (1972) Effect of tillering and cool nights on
photosynshesis and chloroplast starch in Pangola. Crop Sci. 12: 206-208.
Cock J, Yoshida S (1972) Accumulation of C14 labelled carbohydrate before flowering and its subsequent
redistribution and respiration in rice plant. Proc. Crop Sci. Soc. Japan 41 : 226-234.
Eaton F M, Ergle D R (1954) Effects of shade and partial defoliation on carbohydrate level, growth, fruiting
and fiber properties. Plant Physiology. 29: 39-49.
Ernes M J, Wilkins C P, Smith P A, Kupkanchanakul K, Hawker K, Charlton T W A, Cutter E G (1987)
Starch utilization by deep water rice during submergence. In : Proceeding of International Deep Water
Rice Workshop, 26-30 Oct, 1987, Thialand.
Gallagher J N, Biscoe P V, Hunter B (1976) Effect of drought on grain growth. Nature 264 : 541-542.
Hanson A D, Hitz W D (1982) Metabolic responses of mesophytes to plant water deficits. Annu. Rev. Plant
Physiol. 33 : 163- 203.
HiIle Ris Lambers D, Seshu D V (1982) Some ideas on breeding procedures and requirements for deepwater
rice improvement. Page 29-44. In :Proceedings of the 1981 International Deepwater Rice Workshop, IRRI,
Los Banos, Philippines.
Hilliard J H, West S H (1970) Starch accumulation associated with growth reduction at low temperature
in a tropical plant. Science 168 : 494-496.
Hunt L A (1979) Stem weight changes during grain filling in wheat diverse sources. Pages 923-929. In :
Proc. 5th International Wheat Genetics Symposium, Delhi, India.
Jadhav B B (1987) Effect of partial shading on the yield of rice. Indian J. Agric. Sci. 57 (7) : 515-516.
Janardhan K V, Murty K S, Das N B (1980) Effect of lowlight during ripening period on grain yield and
translocation of assimilates in rice varieties. Indian J. Plant Physiol. 23: 163-168.
Levitt J (1980) Excess water or flooding stress. Pages 213-280. In : Response of plants to environmental
stresses, Vol-II Academic press Inc. New York.
Kato K (1986) Effects of the shading and rachis branch clipping on the grain filling process of rice cultivars
differing in the grain size. Japan J. Crop Sci. 56: 173-174.

120

Kemp D R (1981) The growth rate of wheat leaves in relation to the extension zone sugar concentration
manipulated by shading. J. Expt. Biol. 32: 141-160.
Kozlowski T T (1992) Carbohydrate sources and sinks in woody plants. Bot. Rev. 58 (2): 107-222
Kozlowski T T, Keller T (1966) Food relations of woody plants Bot. Rev. 33: 293-382
Lawlor D W, Fock H (1977) Water stress changes in the amounts of some photosynthetic assimilation of
products and respiratory metabolites of sunflower leaves. J. Expt. Bot. 28: 329-337.
Lenhert H J, Vieweg G H, Rongine D E (1979) Starch accumulation in leaves from maize and Zebrina
pendula Schnizl, dependent on nitrogen and phosphorus deficiency. Z. Pflazenphysiol. 92: 163-169.
Marshall B, Sedgley R H, Biscoe P V (1980) Effect of water stress on photosynthesis and respiration in
wheat ears. Aust. J. Agric. Res. 31: 857-871.
Mohanty H K, Khush G S (1985) Diallel analysis of submergence tolerance in rice (Oryza sativa) Theoret.
Appl. Genet. 70 : 467-473.
Morgan J M (1984) Osmoregulation and water stress in higher plants. Annu. Rev. Plant Physiol. 35: 299-319.
Murty KS, Sahu G (1987) Impact of low light stress on growth and yield of rice. Pages 93-101 In :"Weather
and Rice" Proc. International Workshop on Impact of Weather Parameters on Growth and yield of Rice
IRRI, Los Banos, Philippines.
Murty P S S, Rao V R, Charvulu, Murty K S (1983) Relation between spikelet sterility and accumulation
of soluble nitrogen (15N) in rice ( Oryza sativa L). J. Nuclear Agril. Biol 12: 17-19.
Namuco O S, Hsio T C, O' Toole J C (1980) Water stress sensitivity of rice at flowering. Paper presented at the
11th Annual meeting of Crop Science Society of the Philippines, Manila, Philippines.
Nanda J S, Coffman W R (1979) Alternative plant types and other agronomic characteristics for rainfed
lowland rice. In :Rainfed Lowland Rice : Selected papers from the 1978 Internatiional Rice Research
Conference, IRRI, Manila, Philippines.
Palada M C, Vergara R S (1972) Environmental effects on the resistance of rice seedlings to complete
submergence. Crop Sci. 12: 209-212.
Patro R, Sahu A (1988) Effects of low light intensity at different growth stages of crop in sink size. Oryza
23: 123-125.
Perez C M, Palmiano E P, Baun C L, Julliano B O (1971) Starch metabolism in leaf sheaths and culm of
rice. Plant Physiology. 47: 404-408.
Ponnamperuma F N (1984) Effects of flooding on soils. Pages 9-45. In: Flooding and Plant Growth. T. T.
Kozlowski ed. Academic press, London.
Rahman M S, Yoshida S (1985) Effect of water stress on grain filling in rice. Soil Sci. Plant Nutr. 31 (4) :
497-511.
Rai V R S, Murty K S(1979) Note on the effect of complete submergence on RUBP carboxylase activity of
rice seedlings. Indian J. Agric. Res. 13: 61-63.
Rao M J V, Riswas S (1979) Rainfed low land rice in India. In : Rainfed low land rice : Selected papers
from the 1978 International Rice Research Conference. IRRI, Manila, Philippines.
Raakin I, Kende H(l984) Effect of submergence on translocation, starch content and amylolytic activity
in deep water rice. Planta 162: 556-559.
Raysaha P K, HilleRisLambers, Tepora N M (1993) Combination of stem elongation ability with
submergence tolerance in rice. Euphytica 68: 11-16.
Regan K L, Whan B R, Turner N C (1993) Evaluation of chemical dessicants as a selection technique for
drought resistance in dryland wheat breeding programme. Austr. J. Agril. Res. 44: 1683-1691.

121

Reyniers F N, Troung Binh, Jacquinot L, Nicou R (1982) Breeding for drought resistance in dryland rice.
Pages 273-292. In : Drought resistance in crops with emphasis on rice. IRRI, Manila, Philippines.
Setter T L, Waters I, Wallace I, Wiengweera A, Bhekasut P, Greenway H (1989 a) Submergence of rice. I
Growth and photosynthetic response to CO2 enrichment of flood water. Austr. J. Plant Physiol. 16: 251-263.
Setter T L, Jackson M B, Waters I, Wallace I, Greenway H (1989 b) Flood water carbondioxide and ethylene
concentrations as factors in chlorosis development and reduced growth of completely submerged rice. In
: Proceedings of 1987 International Rice Workshop, Thailand.
Singh U S (1982) Deep water rice in India : Problems and prospects. pp 355-362 In : Proceedings of
International Deep Water Rice Workshop, 1981. IRRI, Manila, Philippines.
Trought M C T, Drew M C (1980) The development of waterlogging damage in wheat seedlings (Triticum
aestivum). Shoot and root growth in relation to changes in the concentration of dissolved gases and solutes
in soil solutions. Plant and Soil 54:77-94.
Turner N C, Jones M M (1980) Turgor maintenance by osmotic adjustment : a review and evaluation.
Pages 87-103. In: Adaptation of plants to water and high temperature stress. N C Turner and P J Kramar
eds. John Wiley and sons. London.
Venkateswarlu B (1977) Influence of lowlight intensity on growth and productivity of rice (Oryza sativa
L). Plant and Soil 46: 713-719.
Venkateswarlu B, Srinivasan T E (1978) Influence of lowlight intensity on growth and productivity in
relation to population pressure and varietal reaction in irrigated rice. Indian J. Plant Physiol. 21: (2) 162-170.
Vergara B S (1985) Growth and development of the deep water rice plant. IRRI Research Paper series 103:
1-38.
Wardlaw IF, Eckhardt L (1987) Assimilate movement in lolium and sorghum leaves IV. Photosynthetic
response to reduced translocation and leaf stage. Aust. J. Plant Physiol. 14: 573-591.
Weng J H, Agata W, Takeda T (1986) Studies on dry matter and grain production of rice plants IV. Varietal
differences of total carbohydrate content of heading stage. Japanese J. Crop Sci. 55 (2): 201-207.
Widawsky D A, O' Toole J C (1990) Prioritizing the rice biotechnology research agenda for Eastern India.
Pub. The Rockfeller Foundation.
Yoshida S (1973) Effect of temperature on growth of rice plant ( Oryza satival ) in a controlled environment.
Soil Science Plant Nutr. 19: 299-310.
Yamaguchi T (1973) Studies on the floating rice III. Effects of raising water level on the growth and
carbohydrate contents of the tops. Proc. Crop Sci. Soc. Japan 42: 29-34.
Yamaguchi T, Sato T(1963) Studies on floating rice II. The effect of water level treatment upon carbohydrate
content and the amylase invertase activities. Proc. Crop Sci. Soc. Japan 31: 357-361.

122

STRESS PHYSIOLOGY IN RELATION TO BREEDING

FOR DROUGHT RESISTANCE : A CASE STUDY OF RICE
S. Fukai and M. Cooper 1

SUMMARY
Drought is a major problem for rice grown under rainfed lowland and upland
conditions, but progress in breeding for drought resistance has been slow. This paper
reviews genotypic variation in stress-related traits in rice, and suggests how stress
physiology can contribute to plant breeding programmes that aim to improve yield
under water-limiting environments.
There are several putative traits which confer drought resistance in rice. The
most important is the correct phenology which matches crop growth and
development to the water environment. A deep-root system and some other drought
avoidance mechanisms are useful in upland conditions. However, their usefulness
is not known for rainfed lowland rice where the development of a hard pan prevents
deep root penetration and where transpiration is a small proportion of total water
loss from the system. Drought tolerance mechanisms appear to offer more scope for
the genetic improvement of drought resistance in the rainfed lowland systems.
Osmotic adjustment is promising, as it can counteract the effects of a rapid decline
in leaf water potential and there is large genotypic variation for this trait. Green leaf
retention appears to be a useful character for a prolonged drought, but is affected
by plant size and this complicates the use of this as a selection criterion for drought
resistance.
A major reason for the slow progress in breeding for drought resistance in
rice is the complexity of the drought environment, which often results in the lack of
clear identification of the target environment(s). Then. is a need to identify the relative
importance of the three common drought types; early stress which often causes delay
in transplanting, mild-intermittent stress which can have a severe cumulative effect,
and late stress which affects particularly the late maturing varieties in rainfed lowland rice aerobic and anaerobic soil.

INTRODUCTION
Plant breeding aims to produce new cultivars suitable for a pre- defined target
populations of environments. Often the outcome is improved crop yield by

1 Department of Agriculture, The University of Queensland, Brisbane. Queensland, Australia,

4072

exploiting specific adaptation through minimising adverse effects of biotic factors,

improving broad adaptation across environments or by increasing potential yield
where crop growth is not affected by water or nutrient limitation. Generally
advanced lines from breeding programmes are tested at a number of locations within
the target geographical region for a number of years before their release as new
cultivars. Traditional plant breeding takes time, and strategies to improve the
efficiency of plant breeding programmes are continually sought. One major reason
for the slow progress is the incidence of large genotype by environment (GxE)
interactions, which result from a combination of differences in genotypic adaptation
and the heterogeneous mixture of environments within the target area. The
environmental variability is generally observed as year-to-year and site- to-site
variation. A consequence of GxE interactions is that particular lines often do not
perform well under all conditions encountered in the target population of
environments. The biological reasons for these interactions are often unknown.
Rice is a semi-aquatic plant and commonly grown under flooded
conditions. However, about a half of the rice area in the world not have sufficient
water to maintain flooded conditions and rice yield is reduced to some extent by
drought, a period of no rainfall or irrigation (Hanson et al., 1990). The total area of
upland rice is much smaller than rainfed lowland rice, though drought research has
been concentrated mostly on upland rice, and only recently has a coordinated
research effort for rainfed lowland rice commenced.
A wide range of water stress environments take place in the rice growing area
as drought conditions differ in both the timing and intensity of water deficits.
Therefore, the target population of drought environments for the rice growing area
is a highly complex mixture of different types of water stress environments. In
general the composition of this mixture is poorly understood. However, developing
a picture of this mixture is critical to the success of translating drought physiology
research into successful crop improvement strategies. Stress physiology has
elucidated both responses of different physiological and morphological traits to
water stress and genotypic defferences in these traits under specific drought
conditions. Putative traits have been identified which may confer drought resistance
in some crops, for example, for sorghum (Ludlow and Muchow, 1990). In rice also
a large number of traits have been considered to affect grain yield under drought
conditions (O Toole, 1982). Some traits appear more important than others, and their
usefulness has been confirmed under some drought conditions. However, despite
our increased understranding of stress physiology, development of drought resistant
cultivars, i.e. cultivars which produce higher yield than others in drought conditions,
is slow in rice, as it is for other crops, at least partly because of the lack of our
understanding of GxE interactions.
This paper reviews some of the problems associated with GxE interactions in
physiological studies of drought resistance, examines factors which cause GxE
interactions for rice grown in water-limiting conditions, discusses several putative
traits for improving drought resistance, and suggests ways of improving the
efficiency of plant breeding programmes by adopting physiological research
124

strategies which are appropriate for the target population of environments. The focus
will be on rice, but work from other crops is included to make a comparison with
rice.
G X E INTERACTION AND SELECTION FOR DROUGHT RESISTANCE
Genotype by environment (GxE)interactions are expressed as a change in the
relative performance of genotypes in different environments. These interactions
complicate selection of superior genotypes when they are large relative to genotypic
variation and particularly when they result in a change in the rank of genotypes
across environments (Haldane 1946; Baker, 1988a). The incidence, size and nature
of GxE interactions are commonly investigated by conducting multi-environment
trials (METs), where a sample of genotypes is evaluated at a number of sites in a
number of years. Statistical models and methodology for the analysis and
interpretation of such series of experiments have been widely discussed in the
literature (Comstock and Moll, 1963; Baker, 1988b; Eisemann et al., 1990; Nyquist,
1991; Cooper et al., 1993; Cooper and De Lacy, 1994). Such analysis allows an
evaluation of the size of GxE interaction variation relative to genotypic variation for
the plant traits studied. The objective here is not to review GxE interactions in general
but to consider some aspects of how they can complicate interpretation and
extrapolation from the results of physiological experiments which concentrate on
genotype adaptation to water stress and/or genotype by water-stress (GxE)
interaction in rice grown under water limiting conditions. The two aspects which
will be considered here are (i) definition of targeted environments for the breeding
programme concerned and (ii) the relevance of particular types of physiological
experiments to the target population of environments. The salient principles are
developed by first considering a traditional statistical model used by plant breeders
to analyse the results of genotypic adaptation in METs. This is then re-defined in
terms of environmental factors which impinge on the results of drought resistance
research on rice.
Where genotypes are tested at a series of sites in a number of years, a linear
model which explicitly considers the cross- classification structure of sites and years
is often used to analyse the results of the experiments. This model may be formlly
defined as:
where Pijkl is observation 1 on genotype i at site j in year k and i=l, ..., ng, j=1,
..., ns, k=l, ... ny, l=1, ..., nr with ng , ns , ny and nr the number of genotypes, sites,
years and replicates, respectively; is the grand mean of all observations; gi is the
effect of genotype i; sj is the effect of site j; yk is the effect of year k; (sy)jk, (gs)ij,
(gy)ik, (gsy)ijk are the interaction effects associated with the main-effects; and e ijkl is
the random error effect 1 associated with genotype i at site j in year k. Generally data
on quantitative traits such as grain yield collected from METs are analysed in relation
to this model. From this model the relative importance of genotypic and GxE
interaction effects can be estimated for the traits analysed. Two deficiencies of such
an analysis are (i) that there is no definition of what aspects of the environment are
125

important in discriminating among the genotypes and (ii) there is no understanding

of the physiological attribute to the differences in adaptation expressed among the
genotypes (Eisemann et al., 1990). Consequently where GxE interaction is large many
questions remain unanswered, e.g. were water stress environments sampled in the
METs? What types of water stress environments were sampled? Were the high
yielding genotypes adapted to water stress? How important was GxW interaction?
Was there genotype by water-stress by other (eg. nutrition, disease) environmental
factor interaction?. Clearly these questions will continue to remain unanswered
unless additional information is obtained on the physiological basis of genotypic
variation and the causes of GxE interaction.
While the influence of GxE interactions is an important consideration for some
breeding programmes, the model given in equation (1) is not directly applicable to
many physiological experiments. Often these more detailed experiments concentrate
on one type of water stress environment or a restricted range of water stress
environments. In such experimental studies of drought resistance, we consider that
there is a critical question which must be considered, i.e. how does the type of water
stress generated in the experiment relate to those encountered in the mixture of
environments which comprise the target population of environments?. This question
of relevance can be re-phrased to consider its applicability to specific experimental
situations. We do this below for two types of experiments.
Experiment 1. Where a restricted range of water stress environments is considered
and other environmental factors (eg. nutrition, management, disease, insects etc.)
are held constant, how does this influence the capacity to make inference from the
experiments to the target population of environments? An example of this type of
experiment would be where line-source sprinklers are used to consistently generate
a range of water-stress environments with other inputs held constant (e.g. Puckridge
and O Toole, 1981). Environments at a research station, particularly of high soil
fertility, may be quite different from the target environment where fertilizer input
would be low.
Experiment 2. Where a single specific water stress environment is considered in
addition to the assumption of other environmental factors being held constant, how
does this influence the capacity to make inference from the experiments to the target
population of environments? An example of this type of experiments would be a
specific single environmental screen for drought resistance such as a dry season,
upland experiment using young seedlings in which rainfall occurrence is very low
(e.g. De Datta et al., 1988).
These types of experimental situations are widely encountered in
physiological-genetical studies of drought resistance in rice. However, it is rare for
the questions we have posed to be asked and even more unusual for an answer to
be sought.
To develop a theoretical famework to answer the question of relevance for
experiments 1 and 2 above, we can express the statistical model given in equation
(1) in a form more familiar to the elements of a physiologial experiment. As a first
step, instead of site-to-site and year-to year variation we may consider that different
126

water stress levels (refer to these by the symbol W for water-stress) are encountered
in different experiments. Other environmental factors such as temperature,
nutritional deficiencies and toxicities and biotic stresses (refer to these by the symbol
0 for other) will also influence the performance of the genotypes and these may
vary among experiments. Therefore, the model can be redefined as:
where Pijkl is observation l on genotype i under water-stress j and influenced
by other environmental factor k and i=l, ..., ng, j=1, ..., nw, k=1, ..., no , l=1, ..., nr with
ng, nw , no and nr the number of genotypes, water-stress environments, other
environmental factors and replicates, respectively; m is the grand mean of all
observations; gi is the effect of genotype i; w j is the effect of water-stress type j; ok
is the effect of other environmental factor k; (wo) jk , (gw)ij , (go)ik , (gwo)ijk are the
interaction effects associated with the main-effects; and eijkl is the random error effect
1 associated with genotype i under water- stress j and other environmental facror k.
Clearly this is a simplified model relative to that which would generally apply to
the target production-system. However, it applies to experimental situations where
we can evaluate genotype in a factorial combination of water stress regimes and
other environmental factors. This model is applicable to the sorts of physiological
experiments commonly conducted to investigate the physiological basis of drought
resistance such as our experimental types 1 and 2 above. An important difference
between models (1) and (2) is that developing and analysing series of experiments
in terms of model (2) asks specific questions about the nature of drought resistance
of genotypes in relation to types of water stress. However, to allow such an analysis
it is necessary to be able to identify types of water-stress environments and generate
these within the experiments. A model-based answer to the question of relevance
posed for experiment types 1 and 2 can be obtained by considering the experimental
scenarios posed in context with model (2). This is achieved by placing the appropriate
restrictions on the effects in the model and assessing how the components of model
(2) are confounded (Nyquist, 1991). The two experimental types are considered in
turn.
For experiment 1, where a range of water-stress environments is considered
but other environmental factors are held constant, no =1 and therefore k=1 in model
(2). Here estimates are obtained for genotypic effects across the water-stress regimes
and for GxW interaction. In this situation model (2) becomes:
In this presentation of the model the square brackets surround the effects of
model (2) which are confounded. In this case the estimates of the genotypic effects
are confounded with the genotype by other- environmental-factor (GxO) interactions
and the genotype by water-stress (GxW) interaction effects are confounded with the
genotype by water-stress by other-environmental-factor (GxWxO) interactions.
Therfore, as the GxO and GxWxO interactions increase within the target productionsystem, the capacity to predict the relative drought resistance of genotypes and
expression of GxW interactions in the target production-system from this type of
experiment will decrease. However, it should be emphasised that where GxW
127

interactions are the major component of the GxE interaction complex and both GxO
and GxWxO interactions are small, the results from these types of experiments will
allow good prediction of the relative drought resistance of genotypes in the
environments of the target production-system.
For experiment 2, where only one type of water-stress environment is
considered and other environmental factors are held constant, n o =l and n w =l and
therefore j=k=l in model (2). Here estimates are obtained for genotypic effects only
and no interaction effects can be estimated. In this situation model (2) becomes:
In this case the estimates of the genotypic effects are confounded with all
three components of the GxE interaction complex. Therefore, in this type of
experiment there is no separation of the genotypic effects from the GxE interaction
complex and as any one of the GxE interaction components increases, the
predictive value of the specific environmental screen for the target
production-system will decrease. Again it is important to emphasise that while
prediction to many of the environments which comprise the target production
system may be unsuccessful, if the specific drought resistance screen matches
one type of target-environment in the production-system then the screen will
provide good prediction to that target environment. The additional question
which must be asked in this situation is what is the frequency of occurrence of
the particular target environment. For example, a seedling screen in a dry season
may be useful for identifying rice genotypes which perform relatively well under
early water stress conditions in a wet season, again assuming negligible GxO and
GxWxO interactions. If this form of water-stress occurs with high frequency in
the target production- system then the specific screen would be relevant.
However, if this environment only occurs at a low frequency then while the screen
may have good predictive capacity, its relevance would be decreased.
From the above consideration of screening genotypes for drought resistance
it is argued that a critical component of studying drought resistance is a clear
definition of the relevant target environments within the crop-production system.
Where the objective is to select genotypes or identify physiological components
which contribute to drought resistance, consideration must be given to the relative
contribution of GxW, GxO, and GxWxO interactions in relation to the target
environments. Without due consideration of the relevance of types of experiment,
the success of a particular drought resistance screen, putative drought resistance
trait or plant ideotype is likely to be poor in its contribution to plant breeding
strategies and progress in breeding for drought resistance will continue to be slow.
At present little is known of the size and causes of GxE interactions in rice production
under water limitation. Given the ubiquitous nature of GxE interactions in other
crop production systems (Delacy et al., 1990) and the diverse nature of the types of
environments in which rice is grown it may be anticipated that GxW, GxO and
GxWxO interactions will be large for quantitative traits such as grain yield. In the
following sections we consider putative drought resistance traits which have been
studied in rice and assess their possible contribution to drought resistance for various
128

types of drought conditions. Specific attention is given to the incidence and influence
of GxE interactions on the role of these putative traits in drought resistance.
THE PHYSIOLOGICAL BASIS OF INTERACTING FACTORS
There are a number of reasons for the genotype by drought environment
interactions discussed above. One reason is that yield potential under favourable
water conditions varies among genotypes but yield under water limiting conditions
may be similar. This interaction is not of interest here as we try to understand yield
differences in water limiting conditions.
Timing of drought development in relation to a genotypes phenology is
important for GxE interaction. It is well known that panicle development to
anthesis is the most critical stage for water stress in rice. Our work (Boonjung et
al., 1994b) with rice indicates that grain yield is lowered at the rate of 2% per day
as a 15-day stress period (dawn leaf water potential less than -1.0 MPa) later
during panicle development is delayed (Fig. 1). Unfilled grain number increases
sharply with stress during late panicle development. Water stress reduces
assimilate production and this causes reduction in yield components such as
panicle number per plant, spikelets per panicle, filled grain percentage, and
individual grain mass. However, unfilled grain number appears most susceptible
to reduced assimilate availability.
Assuming the reduction of 2% grain yield per day with the delay in
termination of 15 days stress, a 20-day difference in flowering time between two
cultivars would cause a grain yield difference of about 40% if they have the same
yield potential under non-limiting conditions. Thus, it is likely that the cultivars
with a difference in phenology will react differently to a drought, depending on the
timing of stress development (Maurya and OToole, 1986). Phenology is important
in determining grain yield response also because quick maturing cultivars often
escape from severe stress, while late maturing cultivars may be affected by a terminal
stress. These results suggest that genotypes should be compared for their drought
resistance / susceptibility within the same phenology group, or at least genotypic
variation in phenology should be corrected statistically before genotypic difference
in drought tolerance is estimated (Bidinger et al., 1987a, b). Alternatively it is possible
in some experiments to implement a strategy of staggered planting of different
varieties so that they would flower at about the same time (Lilley and Fukai, 1994b).
In addition to timing of drought in relation to crop phenology, duration and
severity of drought can also affect genotypic response to drought. For example,
genotypic responses to mild, intermittent stress may be quite different from those
to prolonged stress.
Another reason for GxE interactions among drought environments is that
drought affects crop growth indirectly through contributing to the incidence of
another adverse condition. For example, certain diseases in rice tend to develop
under dry conditions. Dry conditions may not always induce the disease in the same
manner. Thus, where there is genotypic variation in resistance to the disease in some
dry years, some lines appears to be doing well relative to others but not in other
129

Fig. 1.

Relationship between grain yield under water stress conditions, relative to the corresponding
sowing in the irrigated trial, and the time of water stress termination expressed in days to
anthesis, of three trials. Experiment 1 (0, Exp 1), Severe stress
SS) and mild stress
(
MS) trials in Experiment 2. S1- S4 indicate time of sowing. (Boonjung et al 1994b).

dry years, depending on the pattern of disease development. Under these situations
the specific drought resistance response is confounded with the presence of the other
factor, a disease in this case, and the large GxWxO interactions (see equations (3a)
and (3b) complicate interpretation of drought resistance responses. It is then
necessary to identify the presence of the disease in the experiment and screening
methods for drought resistance need to consider the complications generated by this
other factor. Where the disease occurs widely with drought, the direct selection
against the disease may be more appropriate than attempting to develop cultivars
with drought resistance per se.
Rainfed lowland rice has a unique nature of alternating aerobic and anaerobic
soil conditions during the crop growth cycle and this is likely to be another major
source of the GxE interactions, particularly GxO and GxWxO interactions. During
a flooded period, rainfed lowland rice may perform similarly to irrigated lowland
rice, but often growth is reduced as water level recedes to below the soil level. Our
recent experience in Northeast Thailand indicates that the rice plants do not show
typical water stress symptoms of leaf rolling and leaf death immediately after the
flood water has disappeared from the soil surface. In some cases these symptoms
were never seen, but grain yield was generally low in the sites where there was no
130

flooding water during the latter part of the growing season. This may be related to
phosphorus becoming less availabe in the aerobic soil, or to a loss of nitrogen during
alternating periods of aerobic and anaerobic soil conditions (Patrick and Wyatt, 1964).
It is also possible that loss of standing water results in increased acidity of soil water,
as in many cases the soil is strongly acidic. With standing water the acid soil is
unlikely to cause a problem. Rice is known to be susceptible to manganese and
aluminium toxicity and reduced pH of the soil solution may induce the toxicity
within a short time period.
Genotypes may differ in their response to the manganese or aluminium level
in the soil as found in some crops. If experiments are conducted at different locations
with soils of various pH levels, there could be strong GxE interactions in dry seasons
and between wet and dry years.
Where the objectives is to study drought resistance, the incidence of the above
complex mixture of these and other types of environment in rice production systems
and their potential for generating GxE interactions strongly suggests (i)the
importance of identifying the types of stress environments sampled in the traditional
METs conducted in plant breeding programme and (ii) more focussed screening of
lines under relevant target environment conditions. In both of the areas there is scope
for stress physiology research to contribute to plant breeding strategies.
PUTATIVE TRAITS FOR IMPROVING DROUGHT RESISTANCE
This section evaluates the evidence for a number of putative traits which may
confer drought resistance in rice. The use of putative traits to improve grain yield
is a classic example of indirect selection in plant breeding. In this case the objective
is to indirectly increase grain yield, or stability of grain yield by selecting for putative
drought resistance traits, or for a combination of yield and the traits. The indirect
response to selection( Gy/x) for character y (yield) from selection for character
x(putative draught resistance trait) is described by the equation :
where ix is the standardised selection differential applied to the putative
drought resistance trait x; hx and hy are the square root of the heritability of the
putative traits x and y, respectively; rgxy is the genetic correlation between traits x
and y; and py is the square root of the phenotypic variation for character y. This
equation may be expressed in different forms and a detailed treatment of the
principles of indirect selection can be found in Falconer (1989) and in the review by
Gallais (1984). The component of equation (4) which determines whether there is
scope for indirect improvement of yield via the putative trait (s) is the genetic
correlation between the two traits. Genetic correlations are a result of either linkage
or pleiotropy. For a high chance of successful indirect selection, identification of a
putative trait which is pleiotropic for yield is desirable. Linkage relationships can
change with time and between populations due to recombination. Recently Aastveit
and Aastveit (1993) considered the impact of GxE interactions on genetic correlations
between traits. In general, the presence of unpredictable GxE interactions will reduce
the heritability of traits and the reliability of genetic correlations between traits
131

measured in individual experiments and thus reduce the indirect response to

selection equation (4).
Where a putative trait is advocated as a selection criterion for a plant breeding
programme it should at the very least be evaluated in relation to the components of
equation (4) in relevant target environments. Information on the heritability of the
trait and the genetic correlation with yield would allow an objective assessment of
the indirect response to selection for yield. The putative trait could be prioritised in
context with the many other traits that are selected for within the breeding
programme. The slow progress which has been observed for the improvement of
grain yield of rice under water limiting conditions suggests that the heritability of
yield is low. In our recent work in Thailand, preliminary information suggets that
genetic variations for yield exists and reasons for the low heritability include large
G x E interactions and experimental errors. Therefore, there appears to be scope for
indirect improvement of yield if the cause of the G x E interations could be better
understood and suitable drought resistance traits could be identified. Here we
consider a number of putative drought resistant traits in rice following a review by
OToole (1982), and evaluate the evidence for these as suitable indirect selection criteria
in relation to the components of equation (4).
Molecular marker technology has been identified as a powerful tool for
selection of traits which are otherwise difficult to screen. OToole (1989) discussed
the use of molecular marker technology for selection of drought resistance traits in
rice. Where molecular markers are linked to genes controlling the expression of
drought resistance traits, there is scope to select for the genes without subjecting
breeding lines to a drought screen. Therefore, given the difficulty in identifying and
conducting a reliable drouhgt resistance screen, there are opportunities for improving
the efficiency of selection for drought resistance via the use of molecular markers.
Work is currently under way in rice and some aspects of this will be discussed.

Root system
Under upland condition, rice, maize and sorghum have similar root length,
density and water extraction pattern down to 60 cm depth (Fukai and Inthapan,
1988). However, below that depth, water extraction by rice is very small and this is
the main reason for the susceptibility of rice to water strees, compared with sorghum
(Inthapan and Fukai, 1988). The develoment of the rice root system is also sensititive
to soil water deficit (Boonjung et al., 1994a).
O, Toole (1982) suggested three root-related adaptive mechanisms for rice. (1)
For relatively large soil water reservoir, increase in rooting depth, root density,
root-shoot ratio and possibly also root conductance. For relatively small soil water
reservoir, (2) possibly increased root penetration of any physical-chemical
impediment and (3) possibly osmotic adjustment of root system. High osmotic
adjustment of roots would allow more thorough extraction of soil water and also
enhanced dehydration tolerance which would enhance root system survival and
plant recovery upon rewatering. However, experimental results to date are not
available to show the effects to osmotic adjustment in roots. Most work on the root
132

system of rice in relation to drought resistance has been conducted under upland
conditions where there is a large soil water reservoir. This relates to OTooles first
point of the adaptive mechanisms, and emphasises the importance of rooting depth,
mot density, and mot thickness.
The development and function of the rice root system, and their relation to
drought resistance were reviwed by Yoshida and Hasegawa (1982). Their review
was mostly based on upland work conducted at IRRI in the 1970s, and their major
findings are summarised here. Rice has a well developed mot system near the soil
surface (0-30 cm), but generally not below 30 cm, compared with other major crops.
There is, however, a lagre variation in root length density below 30 cm among rice
genotypes. Generally genotypes with high root length density below 30 cm have a
deeper root system. Yoshida and Hasegawa used the ratio of deep root weight and
shoot weight as an index for drought resistance (avoidance) as the large deep mot
system would be able to extract more water and the small shoot would require less
water. They found the ratio to be positively related to field evaluation of drought
resistance. Using 1081 entries, they also found that the entries with large deep
root-to-shoot ratios tended to be taller and have less tillers. These characters match
the description of traditional upland rice varieties, since the rice root system is
composed basically of nodal roots, genotypes with large tillers tend to have more
roots, particularly on secondary and tertiary tillers which appear late and have short
mots. Thus, it appears that genotypes with a small number of well-developed tillers
have a smaller number of longer roots and this results in high root length density
at depth. They would be thick (larger diameter) and hence, have high conductance
with large xylem vessels.
While a direct relationship between root length at depth and the amount of
water extracted from the layer is rare. Puckridge and O, Toole (1981) found a deep
rooting cultivar, Kinandang Patong, that extracted a larger amount of water at 40-70
cm depth than the two cultivars IR20 and IR36 which were shallow rooted. Similar
results were shown by Mambani and Lal (1983b, c). Our recent work, also under
upland conditions, (Lilley and Fukai, 1994a) shows that variation in water extraction
among four cultivars was directly related to the variation in root length density.
Thus when the amount of water extracted from a 20 cm layer in the soil profile is
plotted against the root length density in the layer, the data from the four cultivars
fall on one curve for each experiment (Fig. 2). Incresed root length density promoted
the rate of water extraction, though the duration of linear phase of water extraction
was to some extent reduced. In their experiments, ground cover varied among
cultivars, but that apparently did not affect the root length density-extracted water
relationship, indicating the importance of mot length in determining the amount of
water extracted.
Lines with high root length density tend to have high leaf water potential and
delayed leaf death due to water stress (Mambani and Lal, 1983a; Cruz and OToole,
1985; Ekanayake et al., 1985a). These favourable conditions may result in higher grain
yield in lines with high mot length under water limiting conditions(Mambani and
Lal, 1983a), although a yield advantage is not always observed (Puckridge and
133

Fig. 2.

Relationship between a) extractable water, b) maximum extraction rate and c) effective duration
of extraction and root length density in a 0.2 m layer (0.2-0.8m depth) for four cultivars
( CPICB, Lemont, Rikuto Norin 12, Todoroki Wase) in the vegetative stage (solid
symbols) and reproductive stage (hollow) trials. (Lilley and Fukai, 1994a).

OToole, 1981). In the case of Lilley and Fukai(1994b) mentioned earlier, there was
an indication that the cultivar with the highest root length performed better than
others under mild stress conditions, but there was no direct relationship between
total root length and grain yield in the experiment where there was only one period
134

of drought. It is likely that the advantages of varieties with larger root systems would
be greater when there are several drougth periods, i.e. intermittent stress.
However, it is unlikely that the genotypes with high mot length density at
40-80 cm depth will have any advantage under most rainfed lowland conditions, as
the hard pan which develops closer to the soil surface would preclude the expression
of the potentially superior root system. It is rather the genotypes ability to penetrate
the hard pan that may be more important in rainfed lowland conditions. In a lowland
experiment where water stress developed from 45 to 75 days after transplanting,
grain yield of 30 genotypes was not related to any mot characters determined in
aeroponically grown plants (Ingram et al., 1990). Nevertheless in some rainfed
lowland fields where puddling is not practiced, for example due to sandy soil texture,
high root length at depth may have direct relevance.
Root axial resistance can be important in determining the rate of water flow
to the shoot. Rice plants are often not able to extract water thoroughly from deeper
layers, partly because of low root length density at depth, but also increased axial
resistance due to the increased distance to the shoot and also to the smaller diameter
of the roots. Genotypic variation in mot thickness is associated with that in xylem
size (Yambao et al., 1992). However, they consider that increased xylem vessel
diameter will not directly increase the drought resistance of a genotype.
Determination of root length and root depth in the field is time consuming,
and quicker methods have been devised which could be used in screening for an
ideal root system in a breeding programme. For example aeroponic and hydroponic
culture can be used for root system determination (Loresto et al., 1983; Ekanayake
et al., 1985 a; Haque et al., 1989). However, the value of this strategy is questionable
for rainfed lowland rice given the findings of Ingram et al. (1990). Gomathinayagam
et al. (1989) suggest that seminal root growth can bo used for screening deep rooted
rice. In the field the root pulling resistance can be used as a measure of root systems
(O, Toole and Soemenatono, 1981). Plants with high mot pulling resistance were able
to maintain high leaf water potential (Ekanayake et al., 1985b).
Some root system characteristics appear to have high heritability. Using F1,
F2 and F3 populations of the cross between IR20 (a shallow, thin root system) and
MGL-2 (a deep, thick root system), Ekanayake et al. (1985a) found high heritability
for root thickness, mot dry weight and root length density. However, heritability for
root pulling resistance was rather low (Ekanayake et al., 1985b).

Shoot-related mechanisms
OToole (1982) suggested three shoot-related adaptive mechanisms;
accumulation of amino acids or growth regulators, drought avoidance mechanisms
and osmotic adjustment.

Growth regulators
Little work has been conducted with rice in relation to growth regulators and
drought resistance. Dingkuhn et al. (1991a), in the same experiment as will be

135

reported later for osmotic adjustment and other physiological characte of seven
diverse cultivars by Turner et al. (1986a, b), found cultivar differences in abscisic acid
(ABA) accumulation, but the difference was not related to physiological behaviour
of cultivars under water stress. On the other hand, they found stress-induced proline
accumulation to be positively correlated with osmotic adjustment.

Avoidance mechanism
In contrast to a deep root system which can increase water uptake, avoidance
mechanisms in the shoot reduce water use, thus avoiding the development of low
plant water potential.
Epicuticular wax is an avoidance mechanism trait which may have a
significant effect on rice growth during water stress and recovery upon rewatering.
However, the contribution of this putative trait has not been examined in details.
Rice is known to have much less epicuticular wax than other cereals and it has
generally high cuticular (epidermal) conductance (O'Toole et al., 1979). This implies
that rice will lose water even when stomata are completely closed, and this may
lead to rapid leaf death. There is genotypic variation in quantity of epicuticular wax
(OToole, 1982). In our recent experiment, epidermal conductance was found to vary
among rice cultivars, but there was no indication that cultivars with low epidermal
conductance retain green leaves for a longer period.
Leaf diffusive conductance and leaf rolling are also avoidance mechanisms,
for which genotypic variation is well documented (O'Toole and Cruz, 1980). The
variation is at least partly due to differences in the ability to extract water from the
soil, as a result of differences in root system. This causes differences in leaf water
status (water potential), and genotypes which maintain high leaf water potential
tend to maintain high leaf conductance and low leaf rolling(OToo1e and Moya, 1978).
However, the leaf rolling score and leaf water potential relationship varies among
cultivars (Turner et al., 1986 a; Fukai and Inthapan, 1988) possibly because of cultivar
differences in osmotic adjustment. Cultivars with high osmotic adjustment can
maintain turgor potential at relatively high levels under low leaf water potential.
This relationship has been documented in other cereals, though strong experimental
evidence is lacking in rice. Within a genotype, osmotic adjustment is known to affect
leaf rolling in rice (Hsiao et al., 1984).
The direct effect of these drought avoidance mechanisms on growth of rice
plant under limited soil water conditions is not certain, particularly in rainfed
lowland conditions where the amount of water used as transpiration would be a,
rather, small proportion of total water use. The latter includes evaporation from
water or soil surface, deep percolation and seepage. While Dingkuhn et al. (1989a)
found under mild stress conditions a correlation between leaf rolling and water
potential, suggesting that leaf rolling has a positive effect in maintaining high leaf
water potential, our experience is that cultivars which are able to maintain high leaf
water potential have a small degree of leaf rolling. This indicates that, generally leaf
rolling is a result of other avoidance mechanisms, which result in high leaf water
potential.
136

Osmotic adjustment
Osmotic adjustment is an adaptive process in which solutes accumulate in
the cell and osmotic potential is decreased in response to adverse growing conditions
such as drought, salinity and low temperature. As water stress develops osmotic
adjustment is induced, and as such osmotic adjustment increases with decrease in
leaf water potential (Steponkus et al., 1982; Turner et al., 1986 b). Development of
osmotic adjustment however, appears to be rapid in rice, perhaps because of rapid
development of water stress as measured by leaf water potential, compared to
sorghum and maize (Fukai and Inthapan, 1988). Thus any benefits of high adjustment
would be expected in rather early stages of a drought period in rice.
With osmotic adjustment, turgor pressure is maintained at a relatively high
level in spite of reductions in leaf water potential (Cutler et al., 1980). Development
of leaf rolling and half death can be delayed by osmotic adjustment (Hsiao et al.,
1984). It should be pointed out however, that turgor is not necessarily the sole factor
determining the responses of these processes to soil water deficit.
The early work by Steponkus et al. (1982) using four cultivars has shown rather
small variation in osmotic adjustment in rice. The maximum adjustment was about
0.3-0.5 MPa. In the field studies, there was some variation in osmotic adjustment
with a maximum of 0.5 MPa in lowland cultivars and less in upland cultivars among
seven diverse cultivars of rice (Turner el al., 1986b). However, the differences were
related to the development patterns of water stressed cultivars, and the greater
degree of apparent adjustment in the lowland cultivars was due to the greater degree
of cumulative stress. Yang et al. (1983) found some difference in the osmotic potential
water potential relationship between two constrating cultivars, Taichung Native 1
and OS4.
There is no evidance to indicate that genotypic variation in osmotic adjustment
has a positive effect on growth and grain yield in rice, though probably there has
been no serious attempt to test the usefulness of this character in the field. Our recent
work however, does indicate some positive effect of osmotic adjustment in
counteracting low leaf water potential to provide better green leaf area retention
(Henderson et al., 1993). Because osmotic adjustment in rice reaches its maximum
value quickly and the maximum value is maintained for a while, it may be effective
in buffering against the deleterious effect of water stress under situations of rather
a small soil water deficit as may occur under intermittent stress.
The mechanisms of the effect of genotypic variation on yield are well
documented in sorghum, and are briefly reviewed here. Several field experiments
using 2-6 commercial hybrids with a contrast in osmotic adjustment have shown
clear yield advantages of hybrids with high osmotic adjustment when water stress
develops before anthesis or during grain filling (Wright et al., 1983, Ludlow et al.,
1990, Santamaria et nl., 1990). Grain yield under stress conditions relative to that
under well-watered conditions increased linearly with the extent of maximum
osmotic adjustment of the line. Compared to genotypes with low osmotic
adjustment, those with high osmotic adjustment were able to extract water more
thoroughly, maintain higher grain number and translocate a greater amount of
137

pre-anthesis assimilates to fill grains. Using lines from a sorghum population that
had differences in osmotic adjustment, Tangpremsri (1989) showed a yield advantage
of about 0.5 t ha1 in the high osmotic adjustment group. These groups had similar
yield under well watered conditions. Under water-limiting conditions, the high
osmotic adjustment group produced larger maximum leaf area, better leaf retention
during grain filling and more grains per plant. Thus, there are a number of
experimental results to suggest the advantge of high osmotic adjustment when
sorghum experiences a prolonged period of water stress.
The current work by J.M. Lilley and M.M. Ludlow at CSIRO, Brisbane,
Australia, shows a larger genotypic variation in osmotic adjustment than previously
observed in rice. Measurements on rice plants grown in pots in a controlled
environment and subjected to slow (25 day) drying cycle showed that there is a large
range in maximum osmotic adjustment (0.4 - 1.7 MPa) among a set of 59 lines of
diverse adaptation/background. Average osmotic adjustment of lines from upland
environments was lower (0.73 MPa) than that of rainfed lowland, irrigated lowland
and deep water ecotypes (0.96-1.03 MPa). However, there was a large range in
osmotic adjustment within each ecotype.
They also examined variation for osmotic adjustment among a sample of
recombinant inbred lines from a biparental cross. The parents, Morobereken and
CO39, had low (0.7 MPa) and moderate (1.4 MPa) osmotic adjustment, respectively.
Osmotic adjustment of the 52 recombinant inbred lines (RIL-F7), which had been
mapped with RFLP markers, ranged from 0.7-2.7 MPa (Fig. 3). The frequency
distribution of level of asmotic adjustment in the RILs was continuous and there
was no evidence for osmotic adjustment being simply inherited. Transgressive
segregation was observed with 6 lines having significantly higher osmotic
adjustment than the high parent CO39. Morgan (1991) reported that a single recessive
gene accounted for the variation in osmoregulation observed in wheat. Our work
in sorghum (Basnayake et al., 1994a, b) suggests that the inheritance of osmotic
adjustment is at least oligogenic. However, two major genes for high osmotic
adjustment were identified. Current work is investigating the contribution of these
genes to yield under water stress conditions.
The large genotype variation for osmotic adjustment in rice offers considerable
scope for improving these traits in current cultivars. Screening for osmotic
adjustment currently requires intensive measurement procedures and is time
consuming and expensive. Molecular markers such as RFLPs and RAPDs could be
identified for osmotic adjustment genes. The variation in the populations offers a
good possibility of obtaining suitable RFLP markers for these traits so that these
traits could be readily incorporatd in rice breeding programmes.
Water use efficiency
Water use efficiency (WUE) is commonly defined as dry matter production
per unit of water used, and is closly related to transpiration efficiency, dry matter
growth per unit of water transpired. Instantaneous measurements of leaf
photosynthesis and transpiration used to approximate WUE or transpiration
138

Fig. 3.

Frequency distribution of maximum osmotic adjustment of lines from a Co 39 x Moroberekan

cross. (J.M.Lilley and M.M. Ludlow unpublished).

efficiency. Using this technique Dingkuhn et al. (1989b) found WUE to be high in
tropical japonica types, intermediate in indica types and low in aus types, and also
considerable variation in WUE within the types. On the other hand Fukai et al. (1985)
found no significant variation in WUE among four temperate japonica cultivars. They
however, found WUE to decline rapidly with leaf ageing. One of the difficulties of
estimating WUE from instantaneous gas exchanges is the fluctuation of WUE, due
to leaf age or growing conditions such as vapour pressure deficit. This problem can
be overcome by use of stable carbon isotope analysis. Dingkuhn et al., (1991b) found
a good relationship between WUE determined by the gas exchange method and
carbon isotope discrimination. In contrast to drought avoidance mechanisms
expressed in the shoot which decrease the rate of assimilate production, there is no
cost for high WUE, and hence cultivars with high WUE would be advantageous in
dry conditions, unless WUE is associated with characters which are undesirable for
high yield.
Green leaf retention
Since leaf death occurs as a result of severe drought and is an easily
recognisable character in the field, green leaf retention is often used as a selection
139

criterion for drought resistance in rice (De Datta et al., 1988), assuming that cultivars
with good leaf retention produce a higher yield than others in dry environments. In
rice, 30 days old seedlings are subjected to drought and they are visually scored
using a standard (drought score) developed by IRRI. Lines which have better green
leaf retention also recover better after water stress is relieved (Malabuyoc et al., 1985).
Drought at the seedling stage is commonly related to leaf water potential
(Chang et al., 1979). Lines which can maintain high leaf water potential tend to retain
green leaves and hence, have a low drought score. Thus, low drought score generally
indicates that the lines possess mechanisms of drought avoidance, although some
lines appear to be able to maintain green leaves better than others at low leaf water
potential and hence, they are dehydration tolerant (Henderson et al., 1993). Our
recent experiments indicate that genotypic variation in drought score is strongly
associated with variation in light interception prior to the commencement of a stress
period. Thus, lines with large plant size, and hence, with high light interception were
stressed more quickly and had high drought scores. Variation in drought score of a
line is also associated with variation in plant size obtained under various water stress
environments, i.e for some reasons a line may establish more quickly than others in
one site but not in others and this causes inconsistency in relative drought score
among lines. When full ground cover is achieved before drought development, the
genotypic variation in plant size becomes less important in determining leaf death
and this may be a reason for change of genotypic ranking on drought score between
young seedlings and plants with a well developed canopy (Henderson et al., 1993).
In a rainfed lowland experiment where drought was developed in a crop at a high
plant density from 45 to 75 days after transplanting, genotypic variation in drought
score of 30 lines was found to be correlated with yield (r=-0.66) (Ingram et al., 1990).
While leaf death occurs as a result of drought, it does not occur until plants
are severely stressed. Under upland conditions, Boonjung et al. (1994c) have shown
that leaf death is least sensitive to soil water deficit and leaf elongation and leaf
diffusive conductance are much more sensitive. This implies that in rather mild and
intermittent stress where available soil water does not decrease below 50% of the
total available water, plant growth will be affected because of reduced leaf expansion
growth and photosynthesis. Genotypic variation in these characters, rather than
drought score, would determine crop growth. In one of our upland rice experiments
we found no association between leaf conductance and drought score among 20
lines examined, indicating that selection for low drought score may not necessarily
select for high leaf conductance under stress conditions.
Lines with low drought scores do not necessarily perform well in dry
conditions (Puckridge and O Toole, 1981). This again shows the importance of
identifying the target environment and developing an adequate screening method
suitable for the environment. The use of drought score at the seedling stage may be
appropriate for drought that develops early in the season in rainfed lowland; for
example in July in Thailand, particularly for direct seeded rice.
One advantage of lines with good green leaf retention is that they tend to
recover quickly when the stress period terminates. This may be particularly
140

important when stress develops around panicle initiation, as the lines with good
leaf retention can supply more assimilates to the developing panicle. This in turn
will result in production of a large number of spikelets (Lilley and Fukai, 1994b).
Production and retention of fertile spikelets
When a rice variety is grown under different environments, grain yield is
mostly related to grain number as single grain weight is stable across environments.
Grain yield is mostly limited by grain sink capacity to accept assimilates to fill grains
(Fukai et al., 1991). Grain number is determined by the number of spikelets at anthesis
and proportion of spikelets which produce filled grains (filled grain percentage).
The number of spikelets is directly related to the rate of assimilate production
between panicle initiation and anthesis, regardless of whether the assimilate
production is altered by water stress or shading (Boonjung, 1993). Filled grain
percentage on the other hand is related to assimilates produced around anthesis,
and is particularly susceptible to water stress (Cruz and O Toole, 1984, Boonjung,
1993). Maurya and OToole, (1986) found large genotypic variation in filled grain
percentage during the dry season. Reduction in leaf water potential at anthesis may
cause poor exertion of the panicle (Cruz and OToole, 1984; Ekanayake et al., 1989)
and increase the percentage of sterile spikelets because of pollination abnormalities
(Ekanayake et al., 1989, 1990). Ekanayake et al. (1989, 1990) showed some differences
in flowering response to water stress between the two cultivars they studied, and
their work suggests that drought resistant varieties should be able to maintain high
leaf water potential at anthesis and do not show pollination abnormalities as water
potential declines, compared to susceptible varieties.

DEFINITION OF TARGET ENVIRONMENTS

We have argued that it is essential to define the types of drought environment
which are encountered in the target population of environments. It is simply not
sufficient in most cases to say that the breeding programme is designed to improve
drought resistance in rice. The question is, how can this be achieved in a way which
improves the conduct of physiological experiments and selection strategies in plant
breeding programmes? The level of stress encountered in plant breeding experiments
is commonly quantified in terms of the mean yield reduction relative to higher
yielding trials or a high -input control experiment. However, the majority of the
experiments are conducted without knowledge of what type of stress occurred.
Given that different traits confer adaptation to different types of stress, there is a
need to improve on this unenlightened situation to enable assessment of the
relevance of individual experiments to the target environments.
Rice growing environments can be grouped into two major categories, upland
and rainfed lowland conditions. The latter may be further grouped on whether a
hard pan is strongly developed or not. Opportunities exist to characterise the
environment by taking measurements on critical variables such as the depth of free
water in the rainfed paddy. Such physical characterisation is time consuming and
would need to be focussed on critical aspects of the target environments. The
141

objective is not to develop a wish-list of measurable environmental variables but to

streamline the list to allow characterisation of the environment as to its relationship
to the target environments. Another strategy is to use probe genotypes to act as a
bioassay of specific types of stress environments (Eisemann et al., 1990). For example,
near isogenic lines for osmotic adjustment could be used to assay the incidence of
water stress as perceived by the plant.
In characterising an environment where a water deficit develops, it is
important to identify the timing and severity of stress in relation to crop phenology.
Rainfed lowland rice is commonly planted in the monsoon season in Asian countries
where bimodal rainfall is common. There appears to be three types of stress
development in these regions (Chang et al., 1979).
Early stress
Farmers seed in a nursery early in the first rain period and prepare paddies.
In some years however, there is a prolonged dry period between the first rain and
second rain periods. This causes delay in transplanting and reduction in yield
because of the use of old seedlings at transplanting. In some cases the dry period
may occur after transplanting or if direct seeding is practiced, young seedlings may
suffer from water stress. Early maturing cultivars may be affected severely by early
season drought, whereas late maturing cultivars may have sufficient time to recover
from the drought (Maurya and OToole, 1989).
Mild intermittent stress
While the periods of tillering to flowering generally coincide with the highest
rainfall in the monsoon season, short intermittent stress can develop any time during
this period, or any other time. It is known that even in fully flooded paddies, there
can be mid day closure of stomata, probably due to the inability of roots to take up
sufficient water to meet the high water demand. In some crops (e.g. sweet potato,
Bouwkamp, 1989) midday wilting which develops in well-watered soil, is related to
yield reduction among different cultivars. If the water supply is only slightly less
than the demand, there may be no wilting or leaf rolling symptom and hence stress
would be unnoticed by casual observation, but dry matter growth may be affected
and yield reduced by reduced leaf area growth and stomata closure, the plant
responses which are most sensitive to water deficit (Boonjung et al., 1994c).
Late stress
This is a common problem when a late maturing cultivar is used, and growth
during the flowering - grain filling period is affected by drought at the beginning
of the dry season. Early maturing cultivars, or cultivars with appropriate
photoperiod requirement may minimise the end - of - the - season stress. Late stress,
however, is sometimes the result of late planting caused by a dry spell early in the
season.
Identification of the timing and severity of water stress and the magnitude of
the problem are likely to require experimentation for several years. The frequency
142

of occurrence of these three types of water stress environments in the rainfed lowland
regions needs to be determined. Alternatively, if the components of water balance,
particularly seepage and deep percolation, are known for the paddy, a water balance
model may be used to estimate the level of perched water table during crop growth
and this could be used as an indicator of likely plant water stress. Comparison of
grain yield under rainfed conditions and irrigated conditions will provide
information on the magnitude of the drought problem.
It is likely that appropriate traits can be selected once the target environment
is accurately defined. Soil factors are also important in selecting traits, as mentioned
earlier for the root growth, acidity and nutritional problems. Lowered perched water
table in a dry period could interact with these factors, and affect growth indirectly.
Identification of these interactions should help in determining which traits are to be
considered in a plant improvement programme.
The following is a list of suggested putative traits for enhancing the drought
resistance of rice under different water stress conditions. It is emphasized that the
usefulness of each trait for conferring high yield under drought conditions is
unknown in most cases, therefore the traits listed, particularly for lowland
conditions, should be considered as those on which concentrated research effort may
be made by physiologists and plant breeders.
For different cultural conditions
Upland conditions:
Appropriate phenology, deep root system, shoot drought avoidance mechanisms.
Lowland conditions with hard pan:
Appropriate phenology, green leaf retention, tolerance for adverse soil conditions,
high root conductance, root penetration ability.
Lowland conditions without hard pan:
Appropriate phenology, tolerance for adverse soil conditions, deep mot system.
For different types of stress
Early season stress:
Transplanting tolerance, green leaf retention, recovery growth.
Mid season, intermittent stress:
Avoidance mechanisms, appropriate phenology, spikelet retention.
Late season stress:
Appropriate phenology, green leaf retention, tolerance for adverse soil conditions.
Recently crop models have been considered as possible tools to assist
investigation of the physiological and genetic basis of genotype adaptation (Shorter
et al., 1991) and selection decisions in plant breeding (Muchow et al., 1991). Currently
models are not at the level of development where they can be applied to the large
number of unknown genotypes evaluated in breeding programmes. However, they
may be useful in characterising the frequency of the early, mid-intermittent and late
stress environment types in the target population of environments. Our current
143

studies will evaluate the use of a rice model in this role in the rainfed lowland system
in Thailand.

CONCLUSION
Crop physiology can make a strong contribution to a plant breeding
programme designed for improving drought resistance. At present any
improvements in drought resistance are largely a fortuitous outcome of the empirical
multi-environment testing strategies used in the breeding programmes. While this
traditional approach has been effective, its efficiency is widely considered to be low.
We have identified the incidence of strong genotype by water stress (GxW)
interactions for putative drought resistance traits and grain yield as a major
constraint to the genetic improvement of drought resistance. Therefore, plant
breeding strategies and physiological research focussing on drought resistance in
rice must be conducted in a way which allows assessment of the effects of GxW
interactions. We consider that this can be achieved by more effective characterisation
of the type of stress encountered in individual experiments and relating this to
relevant target environments in the rainfed lowland production system. If this
approach is adopted, putative drought resistance traits can be evaluated in context
with the components of indirect response to selection in a way which accommodates
the effects of GxW interactions.
A number of putative drought resistance traits have been identified. The most
important for all drought environments is the correct phenology which matches crop
development to the water availability. Nevertheless, intermittent stress may develop
at flowering time, or the rainy season may finish earlier in some years and the crop
is subjected to stress at flowering. Genotypic variation in sensitivity of flowering to
stress is worth further investigation.
The importance of a deep root system is established as a requirement for rice
cultivars suitable for upland conditions as it ensures greater extraction of water held
in deeper soil profile and therefore, maintenance of a high leaf water potential during
a drought period. However, this is unlikely to be the case for rainfed lowland
conditions, where the existence of hard pan inhibits root penetration to deeper layers.
Water conservation strategies may also have limited success in rainfed lowland
conditions, because of the small quantity of water that can be stored in the soil
compared to the amount of water loss due to deep percolation, seepage and
evaporation from the lowland field. For both upland and rainfed lowland conditions,
cultivars with thick roots are advantageous in maintaining high water flow and
hence, favourable plant water status.
It has been shown that some rice lines have better dehydration tolerance than
others, judging by good green leaf retention even at rather low leaf water potential.
This appears particularly useful for rainfed lowland conditions where probably a
water- stress avoidance strategy has limited value. Dehydration tolerance may be
achieved by increased osmotic adjustment, but this needs further testing. Osmotic
adjustment has been found to vary greatly among rice genotypes.

144

It is essential to demonstrate the value of a trait for dry matter production

during a stress period or a recovery period, or for a direct effect on yield components
such as filled grain percentage or on yield itself, under a drought condition which
is relevant to the target environments. A diagram for this is shown in Figure 4. Once
a positive effect is established in yield determining processes, a screening method
can be devised. Ultimately the success of any such screen will depend on the relative
efficiency of indirect selection for yield via the putative drought resistance trait in
comparison to direct selection for grain yield in the target environments. Therefore,
it is essential that the merit of putative drought resistance traits is evaluated in terms
of the principles of indirect selection.

Fig. 4.

A flow diagram showing processes of the effect of drought on grain yield, and important
factors associated with each process.

It appears also that soil water deficit in rainfed lowlands induces changes in
the environment which are not directly associated with plant water stress but may
affect growth greatly. Soil conditions change from anaerobic to aerobic during cycles
of flooding and drying, and there would be associated changes in nutrient availability
or toxicity. Genotypic variation in response to these changes appears worth further
investigation. Compared with the research effort for drought resistance in upland
rice environments, the research input to the rainfed lowland conditions has been
minimal.
A high priority research area is an accurate definition of the target population
of environments, particularly patterns of drought development. This would assist
definition of optimal multi-envimnment testing strategies for a breeding programme,
ensure determination of the best phenology group for the target environment, and
also focus areas of physiological research which can be directly related to the
objectives of a plant breeding programme.
145

REFERENCES
Aastveit A H, Aastveit, K (1993) Effects of genotype-environment interactions on genetic correlations.
Theoretical and Applied Genetics 86: 1007-1013.
Baker R J (1988a) Differential response to environmental stress. Pages 492-504 In: Proceedings of the second
international conference on quantitative genetics, B S Weir, E J Eisen, M M Goodman and G Namkkoong,
eds. Sunderland, Massachusetts: Sinauer Associates, Inc.
Baker R J (1988b)Analysis of genotype-environmental interactions in crops. ISI Atlas of Science: Animal
and Plant Science 1: 1-4.
Basnayake J, Cooper M, Ludlow M M, Henzell R G (1994a) Inheritance of osmotic adjustment in three
grain sorghum ( Sorghum bicolor, L. Moench) crosses. Theoretical and Applied Genetics (submitted).
Basnayake J, Cooper M, Ludlow M M, Henzell R G (1994b) An investigation of genetic variation for
combining ability of osmotic adjustment among a selected range of grain sorghum lines ( Sorghum bicolor,
L. Moench). Field Crops Res. (submitted).
Bidinger F R, Mahalakshmi V, Rao G D P (1987a) Assessment of drought resistance in pearl millet
[ Pennisetum americanum (L.) Leake]. I. Factors affecting yields under stress. Aust. J. Agric. Res. 38: 37-48.
Bidinger F R, Mahalakshmi V, Rao G D P (1987b) Assessment of drought resistance in pearl millet
[ Pennisetum americanum (L.) Leake]. II. Estimation of genotype response to stress. Aust. J. Agric. Res. 38:
49-59.
Boonjung H (1993) Modelling growth and yield of upland rice under water limiting conditions. Ph D
thesis, The University of Queensland, Australia.
Boonjung H, Fukai S, Henderson S A (1994a) Effects of soil water deficit at different growth stages on
rice growth and yield under upland conditions.I. Radiation interception, water extraction and dry matter
growth under drought. Field Crops Res. (submitted).
Boonjung H, Fukai S, Henderson S A (1994b) Effects of soil water deficit at different growth stages on
rice growth and yield under upland conditions. 2. Phenology, biomass production and yield. Field Crops
Res. (submitted).
Boonjung H, Fukai S, Henderson S A (1994c) Responses of various stress indices to soil water deficit at
different growth stages in rice under upland conditions. Field Crops Res. (submitted).
Bouwkamp J C (1989) Differences in mid-day wilting and yield among sweet potato genotypes. J. Amer
Soc. Hort. Sci. 114: 383-386.
Chang T T, Somrith B, OToole J C (1979) Potential for improving drought resistance in rainfed lowland
rice. Pages 149-164 In: Rainfed Lowland Rice: Selected papers from the 1978 International Rice Research
Conference, International Rice Research Institute, Los Banos, Philippines.
Comstock RE, Moll R H (1963) Genotype-environment interactions. Pages 164-196 In: Statistical genetics
and plant breeding. W D Hansen and H F Robinson, eds. N.A.S.- N.R.C., Publication 982: Washington
D.C.
Cooper M, Delacy I H, Eisemann R L (1993) Recent advances in the study of genotype x environment
interaction and their application to plant breeding. Pages 116-131. In: B C Imrie, and J B Hacker, eds.
Focused Plant Improvenent: Towards responsible and sustainable agriculture. Proceedings Tenth
Australian Plant Breeding Conference, April 1993. Australian Convention and Travel Service, Canberra,
Australia.
Cooper M, DeLacy I H (1994) Relationships among analytical methods used to study genotypic variation
and genotype by environment interaction in plant breeding multi environment experiments. Theoretical
and Applied Genetics (in press).
Cruz R T, OToole J C (1984) Dryland rice response to an irrigation gradient at flowering stage. Agron. j.
76: 178- 182.

146

Cruz R T, OToole J C (1985) Water stress at reproductive stage and grain yield of dryland rice. Philip
Agriculturalist 68: 551-561.
Cutler J M, Shahan K W, Steponkus P L (1980) Alteration of the internal water relations of rice in response
to drought hardening. Crop Sci. 20: 307-310.
De Datta S K, Malabuyoc J A, Aragon E L (1988) A field screening technique for evaluating rice germplasm
for drought tolerance during the vegetative stage. Field Crop Res. 19: 123-134.
De Lacy I H, Eisemann R L, Cooper M (1990) The importance of genotype-by-environment interaction in
regional variety trails. Pages 287-300 In: Genotype-by-environment interaction and plant breeding. M S
Kang ed. Louisiana State University, Baton Rogue, Louisiana.
Dingkuhn M, Cruz R T, OToole J C, Doerffling K (1989a) Net photosynthesis, water use efficiency, leaf
water potential, and leaf rolling as affected by water deficit in tropical upland rice. Aust. J. Agric. Res. 40:
1171-1181.
Dingkuhn M, De Datta S K, Dorffing K, Javellana C (1989b) Varietal differences in leaf water potential,
leaf net CO 2 assimilattion, conductivity and water use efficiency in upland rice. Aust. J. Agric. Res.
40: 1183-1192.
Dingkuhn M, Cruz R T, OToole J C, Turner N C, Doerffling K (1991a) Responses of seven diverse rice
cultivars to water deficits. 3. Accumulation of abscisic acid and proline in relation to leaf water-potential
and osmotic adjustement. Field Crops Res. 27: 103-117.
Dingkuhn M, Farquhar G H, De Datta S K, O Toole I C (1991b) Discrimination of 13c among upland rices
having different water use efficiencies. Aust. J. Agric. Res. 43: 1123-1131.
Eisemann R L, Cooper M, Woodruff D R (1990) Beyond the analytical methodology- better interpretation
of genotype-by environment interaction. Pages 108-717 In: Genotype-by- environment interaction and
plant breeding. MS. Kang, ed. Louisiana State University, Baton Rouge, Louisiana.
Ekanayake I J, De Datta S K, Steponkus P L (1989) Spikelet sterility and flowering response of rice to
water stress at anthesis. Ann. Bot. 63: 257-264.
Ekanayake I J, OToole J C, Garrity D P, Masajo, T M (1985a) Inheritance of root characters and their
relations to drought resistance in rice. Crop Sci. 25: 927-933.
Ekanayake I J, Garrity D P, Masajo T M, OToole I C (1985b) Root pulling resistance in rice: Inheritance
and association with drought tolerance. Euphytica 34: 905-913.
Ekanayake I J, Steponkus P L, De Datta S K (1990) Sensitivity of pollination to water deficits at anthesis
in upland rice. Crop Sci. 30: 310-315.
Falconer D S (1989) Introduction to Quantitative Genetics, Third edition. Longman Scientific and Technical
: New York.
Fukai S, Inthapan P (1988) Growth and yield of rice cultivars under sprinkler irrigation in south-eastern
Queensland. 3 Water extraction and plant water relations-comparison with maize and grain sorghum.
Aust. J. Exp. Agr. 28: 249-252.
Fukai S, Kuroda E, Yamagishii T (1985) Leaf gas exchange of upland and lowland rice cultivars.
Photosynthesis Res. 7: 127-135.
Fukai S, Li L, Vilmonte P T, Fischer K S (1991) Control of grain yield by sink capacity and assimilate
supply in various rice cultivars. Expt. Agri. 27: 127-135.
Gallais A (1984) Use of indirect selection in plant breeding. Pages 45-60. In : Lange, W.; Zeven, A.C. and
Hogenboom, N.G. eds. Efficiency in plant breeding. Proceedings of the 10th Congress of EUCARPIA.
Gomathinayagam P, Natarajan S, Subramanian M, Nagarajan M (1989) A simple procedure to screen deep
rooted rice entries. Oryza 26: 408-410.
Haldane J B S (1946) The interaction of nature and nurture. Ann. Eugenics 13: 197-205.

147

Hanson A D, Peacock W J, Evans L T, Arntzen C J, Khush G S (1990) Drought resistance in rice. Nature
234: 2.
Haque M D, Loresto G C, Chang T T, Tepora N (1989) A comparative study of rice root characters related
to the mechanisms of drought resistance under aeroponic and hydroponic cultures. SABRAO Journal
21: 111-122.
Henderson S A, Fukai S, Lilley J M, George D L, Cooper M, Wamala M H, Watiki J M, Villavicincio J N,
Chinyamakobvu, E, Uaiene R, Ludlow M M (1993) Influence of water stress on leaf death among rice
lines: comparison between glasshouse and field. Pages 220- 223 In: Proceedings of 7th Australian
Agronomy Conference, Adelaide., ETU Publications.
Hsiao T C, OToole J C, Yambao E B, Turner N C (1984) Influence of osmotic adjustment on leaf rolling
and tissue death in rice. Plant Physiology. 75: 338-341.
Ingram K T, Real J G, Maguling M A, Obien M A, Loresto G C (1990) Comparison of selection indices to
screen lowland rice for drought resistance. Euphytica 48: 253-260.
Inthapan P, Fukai S (1988) Growth and yield of rice cultivars under sprinkler irrigation in south-eastern
Queensland. 2. Comperison with maize and grain sorghum under wet and dry conditions. Aust. J. Exp.
Agr. 28: 243-248.
Lilley J M, Fukai S (1994a) Effect of timing and severity of water deficit on four diverse rice cultivars. I.
Rooting patterns and soil water extraction. Field Grops Res. (submitted).
Lilley J M, Fukai S (1994b) Effect of timing and severity of water deficit on four diverse rice cultivars. 3.
Phenological development, crop growth and grain yield. Field Crop Res. (submitted).
Loresto G C, Chaudhary W, Chang T T (1983) Acroponic technique for screening the drought avoidance
mechanism of rice genotypes by the root characters. Garcia de Orta Serie, de Estudos Agonomicos 10: 77-82.
Ludlow M M, Muchow R C (1990) A critical evaluation of the traits for improving crop yields in water
limited environments. Adv. Agr. 43: 107-153.
Ludlow M M, Santamaria F J, Fukai S (1990) Contribution of osmotic adjustment to grain yield of Sorghum
bicolor (L. Moench) under water-limited conditions. II. Water stress after anthesis. Aust. J.Agric. Res. 41:
67-78.
Malabuyoc J A, Aragon E L, De Datta S K (1985) Recovery from drought-induced desiccation at the
vegetative growth stage in direct-seeded rainfed rice. Field Crops Res. 10: 105-112.
Mambani B, Lal R (1983a) Response of upland rice varieties to drought stress. 1. Relation between root
system development and leaf water potential. Plant and Soil 73: 59-72
Mambani B, Lal R (1983b) Response of upland rice varieties to drought stress. 2. Screening rice varieties by
means of variable moisture regimes along a toposequence. Plant and Soil 73: 73-94.
Mambani R, Lal R. (1983c) Responsee of upland rice varieties to drought stress. 3. Estimating root system
configuration from soil moisture data. Plant and Soil 73: 95-104.
Maurya D M, OToole J C (1986) Screening upland rice for drought tolerance. Pages 245-261 In: Progress
in upland rice research. Proceedings of the 1985 Jakarta Conference. IRRI, Los Baos, Laguna, Philippines.
Morgan J M (1991) A gene controlling diiferences in osmoregulation in wheat. Aust. J. Plant Phys.
18: 249-257.
Muchow R D, Hanner G L. Carberry, P S (1991) Optimising crop and cultivar selection in response to
climatic risk. Pages 275-262 In: Climatic risk in crop production: Models and management for the semiarid
tropics and subtropics. R C Muchow and J A Bellamy Eds. C.A.B. International: Wallingford.
Nyquist W E (1991) Estimation of heritability and prediction of selection response in plant populations.
Crit. Rev. Plant Sci. 10: 235-322.

148

OToole J C (1982) Adaptation of rice to drought-prone environments. Pages 195-213 In: Drought resistance
in crops with emphasis on rice. IRRI, Los Banos, Laguna, Philippines.
OToole J C (1989) Breeding for drought resistance in cereals: Emerging new technologies. Pages 81-94 In
: Drought resistance in cereals. F W G Baker Editor, CAB International, Wall-ingford, Oxon, UK.
OToole J C, Cruz R T, (1980) Response of leaf water potential, stomatal resistance, and leaf rolling to
water stress. Plant Physiology. 65: 428-435.
OToole J C, Cruz R T Seiber J N (1979) Epicuticular wax and cuticular resistance in rice. Plant Physiology.
47:239-244.
OToole J C, Moya TB (1978) Genotypic variation in maintenace of leaf potential in rice. Crop Sci. 8:873-876.
OToole J C, Soemaratono (1981) Evaluation of a simple technique for characterizing rice root systems in
relation to drought resistance. Euphytica 30: 283-290.
Patrick W H Jr, Wyatt R (1964) Soil nitrogen loss as a result of alternate submergenceand drying. Soil Sci.
Soc. Am. Proc. 28:647- 653.
Puckridge D W, OToole J C (1981) Dry matter and grain production of rice, using a line source sprinkler
in drought studies. Field Crops Res. 3:303-319.
Santamaria F J, Ludlow M M, Fukai S (1990) Contribution of osmotic adjustment to grain yield in Sorghum
bicolor (L. Moench) under water-limited conditions. I. Water stress before anthesis. Aust. J. Agric. Res.
41:51-65.
Shorter R, Lawn R J, Hanner G L (1991) Improving genotypic adaptation in crops-a role for breeders,
physiologists and modellers. Expt. Agri. 27 :155-175.
Steponkus P L, Shahan K W, Cutler J M (1982) Osmotic adjustment in rice. Pages 181-194 In: Drought
resistance in crops with emphasis on rice. IRRI, Los Banos, Laguna, Philippines
Tangpremsri T (1989) Physiological studies of the performance of sorghum genotypes under moisture
limiting conditions. Ph D Thesis, The University of Queensland, Australia.
Turner N C, OToole J C, Cruz R T, Namuco O S, Ahmad S (1986a) Responses of seven diverse rice cultivars
to water deficits. 1. Stress development, canopy temperature, leaf rolling and growth. Field Crops Res.
13:257-271.
Turner N C, OToole J C, Cruz R T, Yambao E B, Ahmad S, Namuco, O S, Dingkuhn, M (1986b) Responses
of seven diverse rice cultivars to water deficits. 2. Osmotic adjustment, leaf elasticity, leaf extension, leaf
death, stomatal conductance and photosynthesis. Fields Crops. Res. 13:273-286.
Yambao E G, lngram K T, Real J G (1992) Root xylem influence on the water relations and drought
resistance of rice. J. Expt. Bot. 43 :925-932.
Yang C M, Chen C L, Sung F J M (1983) Water relations in rice: 1.Leaf diffusive resistance and osmotic
adjustment of field-grown rice genotypes. Proceedings of the National Science Council, Part B, 7:288-292.
Yoshida S, Hasegawa S (1982) The rice root system: its development and function. Pages 97-114 In:
Drought resistance in crops with emphasis on rice. IRRI, Los Banos, Laguna, Philippines.
Wright G C, Smith R C G, McWilliam J R (1983) Differences between two grain sorghum genotypes in
adaptation to drought stress. l. Crop growth and yield responses.. Aust. J. Agric. Res. 34:615-666

149

VARIETAL CHARACTERISTICS AND SOIL

MANAGEMENT TO REDUCE DROUGHT STRESS
H.U. Ahmed, M.L. Ali, S.K. Zaman,
M.A. Kabir and N.M. Miah 1

SUMMARY
The Rainfed Lowland Rice Research Consortium site in Bangladesh is located
at northwest part, known as Barind Tract, which is characterized by low annual
rainfall. Rice crop in this region often faces drought stress at reproductive phase.
Soil is poor in organic matter with hard and thick (8.0 cm) ploughpan. Two hundred
and fortyeight advanced breeding lines and varieties were screened and evaluated
with respect to different plant characters under field condition to find out suitable
varieties which can tolerate drought at later growth stages. Fortytwo advanced
breeding lines were identified as drought tolerant, out of which two
lines-BR4974-45-9-2 and BR4974-42-1-3 have deep rooting ability and can even
penetrate such hard ploughpan and also produce higher grain yield than the local
varieties under rainfed condition. Another line, IR41054-81-2-3-22 identified as early
maturing was found to escape drought and produce about 30% more grain yield
than the local popular early maturing variety. Soil management manipulation
treatment for reducing the ploughpan thickness through deep tillage and ploughpan
perforation through dhaincha (Sesbania aculeata ) roots significantly increased grain
yield and root mass of rice genotypes.

INTRODUCTION
About 36% of South and Southeast Asian rice growing areas come under
rainfed lowland rice ecosystem (BRRI, 1992). Uncertainty characterizes rice farming
in this ecosystem where it often suffers from drought, floods, pests, weeds and soil
constraints, sometimes all in the same field and in the same season. Rainfall is erratic
and other conditions for growing rice crop are diverse and unpredictable.
In Bangladesh, about 44% of total rice area is occupied by rainfed lowland
rice (transplant aman), out of which about 50% is covered by high yielding varieties
(HYV) developed by Bangladesh Rice Research Institute (BBS, 1992). Rest of the
areas are still being covered with traditional varieties, characterized by low yield
and photoperiod sensitivity enabling for late planting due to late recession of flood
water and late harvesting of preceding aus and jute crops.
The existing HYV are not suitable for cultivation under unfavourable rainfed
lowland situation and may suffer from drought, low temperature and soil stress at
1 Bangladesh Rice Research Institute, Gazipur, Bangladesh

reproductive and ripening phases particularly when the crop is late planted.
Therefore, steps are to be taken to develop varieties and associated technologies to
overcome drought, low temperature and soil stresses.
Drought in rice is a real concern and a serious yield limiting factor on much
of the worlds 70 million hectares of non- irrigated rice (OToole and Chang, 1978).
Lack of rainfall and the resultant soil moisture deficit form the basis of drought. TO
overcome drought stresses, plants develop some defensive mechanism such as
stomatal closure, thick cuticle, leaf rolling, profuse, or deep and thick root system
(OToole and Chang, 1978). Early maturity is also a way to escape drought stresses.
Extensive study is difficult to test all the above characters for a large number
of varieties/lines. Field screening by visual scoring would appear to be an indirect
means of screening for root system development (OToole and Chang, 1978). The
percentage of filled grain on the panicle is a useful criteria in ascertaining reactions
to reproductive phase drought. Proportional reduction in yield in a genotype due
to stress acts as a indicator of drought resistance or susceptibilitiy (OToole and
Chang, 1978). Plant height and panicle exertion is also an indication to resistance or
susceptibility to drought. Root system plays an important role under water deficit
conditions and the nature and extent of root development are major factors
governing plants response to moisture condition (Russell, 1959). Haque et al.,
(1988) also stated that selection based on root characters for drought resistance could
be done in the early stage of growth. A thick and dense ploughpan layer restricts
the rootzone volume of soil and hence reduces its moisture storage capacity.
The rainfed lowland rice research consortium site located in the Barind Tract
has a ploughpan at shallow depth (about 10 cm) which creates adverse conditions
for root development of rice crop. The Barind tract receives less annual rainfall than
most of the parts of Bangladesh (Fig. 1).
The distribution of rainfall is concentrated in a few months, from June to
September (Fig.2).
Transplant aman often faces terminal drought stress. The soil in this area is
also low in organic matter content, which adversely affects other soil properties
(Table 1). Rainfall is the only source of water for crop growth because large scale
irrigation is not possible due to undulating land scape and poor aquifers. To address
some of these problems a number of experiments were conducted at the site from
1991 to 1993.
Experiment 1. Screening of advanced breeding lines for drought tolerance at
reproductive phase under rainfed condition.

Materials and methods

Eighty seven advanced breeding lines/varieties in 1991, 62 in 1992 and 99 in
1993 were grown for screening reproductive phase drought tolerance. In 1991, twenty
eight-day old seedlings were transplanted in two different sites- high and lowland
situation at Alimganj and in irrigated plot at BRRI, Rajshahi station. Each entry had

151

Fig. 1.

Map of Bangladesh Showing Barind area and mean annual rainfall.

5m x 2 rows plot with three replications in RCB design having normal and late
planting in all the three situations. In 1992, thirty-day old seedlings were transplanted
in irrigated field and forty three-day old seedlings were transplanted in rainfed field
each with 5m x 2 rows plot in three replications in RCB design only at BRRI, Rajshahi
152

Fig. 2.

Monthly rainfall, Alimganj, 1991-93

station. In 1993, thirty-day old seedlings were transplanted under rainfed condition
in 4m x 2 rows plot at Alimganj with three replications following RCB design. The
seeds were sown in late July to face the drought stress. Two-three seedlings were
planted per hill at a spacing of 25 x 15 cm with a fertilizer dose of 80:60:40 kg NPK/ha
for all three years. Total quantity of phosphorus and potassium were applied as
basal. Nitrogen was applied in two equal splits- one at tillering stage and another
at panicle initiation stage. Insect pests were controlled as and when necessary. Soil
moisture content from 7th October to harvesting stage was also recorded (Fig. 3).

153

Fig. 3.

Table 1

Soil moisture content, Alimganj, 1993

Physical and chemical properties of the experimental field at Alimganj.

T.Aman, 1993.

Textural Class
Sand (5%)
Silt (%)
Clay (%)
Water holding capacity (%)
Bulk density (g/cm3)
Soil pH
Organic matter (%)
Total N (%)
Available P (ppm)
Exchangeable K (meq/100g)
Available S (ppm)
Available Zn (ppm)
(01NHC1)
Cation exchange capacity
(meq / 100 g)

154

Silty Clay
5.00
51.00
44.00
57.00
1.58
5.8
1.48
0.056
15.00
0.23
15.00
2.3
15.00

Results and discussion

In 1991, twelve lines were selected on the basis of panicle exertion ratio and
spikelet sterility percentage (Table 2a).
Table 2a.

List of the advanced breeding lines selected for drought tolerance at

reproductive phase. T. Aman, 1991.
Pancile exertion ratio

Advanced lines

Highland

Lowland

Sterility percentage

Irrigated

Highland

Lowland

Irrigated

1st

2nd

1st

2nd

1st

2nd

1st

2nd

1st

2nd

1st

2nd

set

set

set

set

set

set

set

set

set

set

set

set

BR4548-2B-70

0.955 0.869

1.005

0.885 1.010 0.946 20.24 34.17 19.24 31.72 20.17 28.39

BR2575-2B-62-1-1

1.022 0.928

1.088

0.868 1.155 0.937 38.49 69.95 23.37 42.30

28.34 71.00

BR4970-37-1-2

0.991 0.882

0.935

0.898 1.127 0.990 35.17 56.52 24.43 49.02

28.48 63.52

BR4970-107-2-6

1.036 0.960

1.094

1.033 1.058 1.138 35.42 55.12 36.34 45.40

34.79 39.37
22.10 66.66

BR4974-23-1-1

1.066 0.874

1.194

0.932 1.176 0.895 31.11 39.39 27.17 47.33

BR4974-45-9-2

1.079 0.874

1.188

0.999 1.261 1.051 39.52 40.20 44.45 41.20

3211 32.56

BR4548-2B-222

0.825 0.913

1.058

1.023 1.168 1.004 45.17 58.04 34.30 43.67

16.93 38.66

BR5437-2B-13-2

1.034 0.947

1.038

0.946 0.984 0.921 10.49 21.23 15.08 22.10

13.13 42.80

BR4974-13-2-9

0.884 0.862

0.941

0.896 1.104 0.805 34.92 72.48 44.42 54.61

21.19 97.28

BR5204-B-24

0.994 0.983

0.974

0.730 0.999 0.846 19.26 41.65 14.43 56.34

17.72 46.70

BR5323-B-52

1.055 0.843

1.075

0.944 1.104 0.930 27.32 41.82 20.62 31.18

16.24 25.22

BR1870-88-1-1

0.947 1.015

0.943

0.996 0.977 0.854 34.86 41.97 37.91 28.62

16.40 28.69

BR11

0.975

0.950

0.938

44.49

24.84

26.86

BR22

0.894

0.833

0.873

48.19

54.58

22.06

Nizersail

0.959

1.066

0.974

43.62

34.68

16.47

Date of Seeding:
Date of Transplanting :

Month

Normal planting

Late planting

17.7.91
14.8.91

6.8.91
17.9.91

Temperature
Minimum (range)c

Maximum (range)c

October

19.0-25.5

25.0-35.0

November

10.0-19.0

26.0-30.5

December

7.5-13.5

17.2-28.0

155

All the twelve entries showed good panicle exertion both in the normal (1st
set) and late planting (2nd set) in case of high land and low land situation compared
to irrigated plot. The spikelet sterility of all the twelve entries of the normal planting
(1st set) under high land, lowland and irrigated situations was lower compared to
check entries. BR2575-2B-62-1-1, BR4974-23-1-1, BR5437-2B-13-2 and BR4974-13-2-9
showed exceptionally high spikelet sterility in the irrigated situation.
In 1992, twelve lines were selected on the basis of visual panicle exertion and
visual fertility score under field condition (Table 2b).
Table 2b.

List of the advanced breeding lines

reproductive phase, T.Aman, 1992.
Plant height

Advanced

(cm)

Date of

selected

for drought tolerance at

Panicle exertion

flowering

score

Fertility
score

lines/
varieties

Irrigated Rainfed

Irrigated Rainfed

Irrigated Rainfed

Irrigated Rainfed

BR4761-3B-7-2

104

62

14/11

16/11

BR4761-3B-108-2

112

61

14/11

12/11

BR4764-3B-3-2

126

72

12/11

14/11

BR4972-92-1-3-1

100

56

12/11

14/11

BR4972-92-1-3-1

100

52

12/11

14/11

BR4761-3B-208-1

100

61

12/11

14/11

BR4761-3B-18-4

106

62

13/11

14/11

BR4961-3B-70

113

85

8/11

15/11

BR1192-2B-10-1-1

86

58

BR1867-26-3-1-1

92

63

14/11
13/11

16/11
16/11

1
3

3
3

1
1

5
3

BR850-22-1-4

88

68

10/11

82

17/11

3
2

103

11/11
18/11

BR4974-48-2-3

Nizersail (ck.)

94

78

12/11

14/11

BR22 (ck.)
BR23 (ck.)

85
83

63
51

12/11
14/11

14/31
16/11

3
5

3
5

1
5

3
5

Date of seeding 23-07-92

Date of transplanting Irrigated = 26-8-92; Rainfed = 08-9-92

Plant height and date of flowering were also considered. All the selected
entries showed fertility score of 1, except BR4961-3B-70 and BR974-48-2-3 which
showed score of 3 and 2, respectively, under irrigated conditions. However,
fertility score under rainfed condition varied between 3 and 5 for all entries
including checks.
156

Table 2c.

List of the advanced breeding lines and varieties selected for drought
tolerance at reproductive phase. T. Aman, 1993.
Plant

Date

Panicle

height

of

exertion

(cm)

flowering

score

BR4872-72-1-4-3-1

128.2

11/11

BR4872-72-1-4-3-3

141.6

11/11

BR5226-21-2-3-2

126.6

6/11

BR5296-29-1-7-2

119.2

16/10

BR5439-39-8-1

119.7

23/10

BR5800-9-1

124.2

31/10

BR1725-13-7-1-6

105.4

21/10

BR4974-2-6-1-3-1

143.2

30/10

BR5204-213-2-9

128.4

27/10

BR5226-21-6-3-3

122.2

30/10

BR5225-2-3-1-4

120.1

20/10

BR5323-B-73-2-2

141.8

25/10

BR5452-6-2-2

123.4

10/ 10

BR5455-16-1-1

130.4

22/ 10

BR5479-11-6-1

119.1

26/10

BR554-156-1-2-1-1-519

103.1

2/11

BR5226-5-4

125.0

3/11

BR830-22-1-4

106.4

2/11

BR11(ck.)

106.3

23/10

BR22(ck.)

114.3

7/11

Batraj (ck.)

130.4

3/11

Advanced lines/
varieties

Date of sowing : 3.7.93

Fertility
score

Date of Transplanting : 25.7.93

In 1993, 18 lines were selected on the basis of visual panicle exertion and visual
fertility score. List of the selected entries is presented in Table 2c. Most of the entries
flowered earlier than BR22 (photoperiod sensitive modern variety). However, only
three entries BR4872-72-1-4-3-1, BR4872- 72-1-4-3-3 and BR-5226-21-2-3-1 flowered
at more or less same time with BR22. These three lines had better panicle exertion
and fertility score than BR22 and assumed to have some degree of tolerance to
drought stress at reproductive phase.
157

Experiment 2. Evaluation of selected advanced breeding lines for reproductive

phase drought tolerance.

Materials and methods

Eighteen advanced breeding lines selected from the field screening for
reproductive phase drought tolerance under rainfed condition in 1991 and 1992 wet
season and three popular transplant aman varieties were included in this trial.
Twenty five-day old seedlings were transplanted in a plot of 5m x 6 rows with a
spacing of 25 x 15 cm using 2-3 seedlings per hill. Fertilizers were applied @ 80:60:40
kg NPK/ha. Total quantity of phosphorus and potassium was applied as basal and
nitrogen was applied in two equal splits, one at tillering stage and another at panicle
initiation stage as top dress. Insect-pests were controlled as and when necessary. The
seeds were sown on 3rd July and transplanted on 28th July under both rainfed and
irrigated conditions at Alimganj.

Fig. 4.

158

Soil moisture content, Alimganj, 1993

Observations were recorded on plant height, phenotypic acceptability, date of

flowering, panicle exertion, spikelet sterility, plot yield and root mass density upto
30 cm with an increment of 10 cm. Soil moisture content from 7th October to
harvesting stage was also recorded (Fig. 4).

Results and discussion

Out of eighteen advanced lines, two lines viz. BR4974-42-1-3 and
BR4974-45-9-2 were identified as promising with respect to tolerance to drought
stress. These two lines showed better panicle exertion and lower spikelet sterility
and had comparable yield with checks under irrigated condition and better yield
under rainfed condition (Table 3). In respect of root mass density distribution,
BR4947-42-1-3 had 3.78 % root within 10-20 cm depth and 0.81% below 20 cm
keeping 95.41% root within 0-10 cm and BR4997-45-9-2 had 92.37% root at 0-10 cm
depth, 6.22% at 10-20 cm and 1.41% at 20-30 cm soil depth (Table 4).
BR4764-3B-3-2, BR4761-3B-7, BR2575-2B-62-1-1, BR4970-47-1-6 and BR497423-1-1 have also good distribution of toot below 10 cm depth (Table 4) and presumed
to have some degree of drought tolerance. Though the root distribution of
BR4761-3B-7 was not so good but it gave good yield (3.35t/ha) under rainfed
condition presumably due to physiological characters such as stomatal closure, thick
cuticle, leaf rolling etc (OToole and Chang, 1978).

Experiment 3. Evaluation of photoperiod insensitive advanced breeding lines for

early maturity.

Materials and methods

In 1992, eleven advanced breeding lines and one check (BR14) were grown in
lowland situation at Alimganj. The experiment was conducted in RCB design with
3 replications. The plot size was 5m x 8 rows with spacing of 25 x 15 cm. In 1993,
three advanced breeding lines were tested for early maturity. The experiment was
conducted in RCB design with 3 replications under strictly rainfed condition at
Alimganj in three different toposequences viz. high, medium and low. The plot size
was 5m x12 rows in each toposequence with a spacing of 25 x 15 cm. Fertilizer dose
was 80:60:40 kg NPK/ha for both the years. All phosphorus and potassium were
applied as basal and nitrogen in two equal splits as top dress, one at tillering stage
and another at panicle initiation stage.

Results and discussion

In 1992, out of 11 advanced lines, three lines namely, BR5363-10-2-1, BR5363-10-2-3
and IR41054-81-2-3-2 were selected on the basis of higher grain yield and early maturity
(Table 5). These lines also matured earlier than the check variety, BR14.
The growth duration of all the three advanced lines ranged from 117 to 125
days in 1993 (Table 6). Varieties with such growth durations can easily escape drought
and can be harvested in early November. Though yield differences between the
toposequences were low, yet on the basis of yield, and phenotypic acceptability at
159

maturity and growth duration compared to the checks (BR14 and Biharibatraj),
IR54104-81-2-3-2 with an average yield of 3.83 t/ha having long slender grain was
selected which can substitute BR14 in this region.
Table 3.

Grain yield and some ancillary characters of advanced breeding lines

selected for reproductive phase drought tolerance grown in T. Aman,
1993 at Alimganj.

Advanced

Plant

Date

Pancile

Spikelet

Yield

Phenotype

height

of

exertion

sterility

(t/ha)

acceptability

(cm)

flowering

ratio

lines

Irrig- Rainated

Irrig-

Rain- Irrig-

fed

ated

fed

at maturity

Rain- Irrig-

Rain- Irrig-

Rain-

Irrig- Rain-

ated

fed

ated

fed

ated

fed

ated

fed

BR4761-3B-7

145

138

8/11

8/11

1.02

1.01

21

38

4.06

3.35

BR4761-3B-7

145

138

8/11

8/11

1.02

1.01

21

38

4.06

3.35

BR4761-3B-108-2

112

110 30 / 10

30 /10

1.01

1.01

17

20

2.93

2.75

BR4761-3B-208-1

121

99

8/11

9/11

1.14

1.11

21

31

3.46

2.70

BR4761-3B-18-4

127

104

7/11

8/11

1.06

1.04

30

45

3.33

2.30

BR4764-3B-3-2

140

118

9/11

10/11

1.08

1.03

23

29

2.74

2.18

BR4972-92-1-3-1

124

113

6/11

5/11

1.10

1.07

24

31

3.25

2.63

BR4761-92-1-3-1

132

115

6/11

5/11

1.10

1.08

34

33

2.93

2.50

BR4761-38-70

149

124

7/11

6/11

1.06

1.05

36

25

2.27

2.12

BR1192-2B-101-1

148

122

9/11

6/11

1.03

1.01

29

37

2.88

2.43

BR1867-26-3-1-1

135

118

4/11

2/11

1.02

0.97

30

46

3.20

2.12

BR850-22-1-4

151

121

4/11

31/10

0.97

0.94

26

33

3.17

1.95

BR4974-48-2-3

131

110

7/11

4/11

1.02

0.99

22

26

3.55

3.50

BR2575-2B-62-1

117

105

7/11

4/11

1.07

1.06

24

27

3.66

2.13

BR4970-47-1-6

118

104

6/11

5/11

1.02

1.02

28

37

3.60

3.25

BR4974-23-1-1

140

115

3/11

4/11

1.06

1.03

34

28

3.55

2.95

BR4974-34-6-6

149

117

5/11

5/11

1.03

0.98

31

36

3.71

3.00

BR4974-42-1-3

154

117

4/11

4/11

1.03

1.01

27

32

4.26

3.48

BR4974-45-9-2

148

114

4/11

5/11

1.11

1.05

31

27

3.87

3.46

BR11 (ck.)

111

95 24/10

20/10

0.96

0.97

28

28

4.34

3.25

BR22 (ck.)

122

101

10/11

7/11

0.93

0.94

28

29

4.19

2.95

Niwrsail (L.ck.)

144

125

8/11

4/11

1.01

1.09

17

29

3.17

2.86

Date of sowing: 3.7.93

160

Date of Transplanting : 28.7.93

Table 4.

Root weight and distribution pattern of some advanced lines selected

for reproductive drought tolerance grown under rainfed condition, T.
Aman, 1993 at Alimganj.
Rootmass weight (mg/cm3)

Distribution (%)

Soil layer (cm)

Soil layer (cm)

Advanced lines

0-10

10-20

20-30

0-10

10-20

20-30

BR4761-3B-7

2.364

.033

.012

98.13

1.37

0.50

BR4761-3B-20-2

2.418

.041

.014

97.66

1.65

0.69

BR4761-3B-208-1

2.447

.059

.034

96.34

2.32

1.34

BR4761-3B-18-4

3.086

.049

.019

97.85

1.55

0.60

95.74

2.96

1.30

BR4764-3B-3-2

2.810

.087

.038

BR4972-92-1-3-1

2.345

.059

.024

96.58

2.43

0.99

BR4772-92-1-3-2

2.466

.044

.019

97.51

1.74

0.75

BR4761-3B-70

2.412

.095

.022

95.37

3.76

0.87

BR1192-2B-10-1-1

2.823

.067

.028

96.74

2.30

0.96

BRl867-26-3-1-1

2.973

.077

.022

96.78

2.51

0.71

BR850-22-1-4

3.115

.059

.016

97.65

1.85

0.50

BR4974-48-2-3

2.842

.051

.027

97.33

1.75

BR2575-2B-62-1-1

2.728

.075

.038

96.02

2.64

1.34

BR4970-47-1-6

2.435

.085

.019

95.90

3.35

0.75

BR4974-23-1-1

2.356

.063

.039

95.85

2.56

1.59

BR4974-34-6-6

2.772

.059

.038

96.62

2.06

1.32

.079

.017

95.41

3.78

0.81

92.37

BR4974-42-1-3

1.994

0.92

BR4974-45-9-2

1.961

.132

.030

6.22

1.41

BRll (ck.)

2.002

.063

.020

96.02

3.02

0.96

BR22 (ck.)

3.133

.072

.012

97.39

2.24

0.37

Nizersail (ck.)

2.607

.057

.050

96.06

2.10

1.84

Experiment 4. Ploughing layer, ploughpan thickness, soil moisture and root

distribution.
Materials and methods
A survey was done during 1st and 2nd week of November in 1991 when the
rice crop was in milk to dough stage in Alimganj area to investigate the ploughpan
depth and thickness in relation to root distribution at high and low toposequences.
161

Due to undulating landscape, sampling was done from high and low toposequences
dividing each into 3 sites and from each site 10 sampling spots were chosen.
Table 5.

Grain yield and some ancillary characters of the early maturing

photoperiod insensitive breeding lines tested in Alimganj during
T.Aman, 1992.

Advanced lines

Plant
height

Panicle
m/ 2

(cm)

Phenotypic
acceptability at

Days
to

Days
to

Grain
yield

Veg.

Mat.

flow.

mat.

(t/ha)

BR5000-2B-6-1

143

188

100

127

2.0

BR5344-38-2-2

121

178

112

139

2.3

BR5363-10-2-1

105

244

92

119

3.6

BR5363-10-2-3

112

256

92

119

3.7

IR28941-1-3-5-1-2

82

224

106

132

2.9

IR38499-60-368-3-21-1

95

177

95

121

2.7

IR40931-26-3-3-5

106

194

107

133

3.2

IR41054-81-2-3-2

97

260

93

119

3.7

IR48776-10-2-1

104

246

93

120

2.3

B3894-40D-PN-5-1

109

204

97

124

2.9

B5986-MR-B-1-10

99

234

94

121

3.1

102

234

96

124

3.8

BR14(ck.)
Date of sowing : 10-7-92;

Date of transplanting : 9-8-92

Results and discussion

The study revealed that the ploughpan depth ranged from 12.5 to 17.5
cm. The ploughpan thickness was 7.5 cm and 10.0 cm for low and high
toposequences, respectively. On an average, maximum root weight ranged from
5.9-9.5 mg/cm3 and major root distributions were found within 0-10 cm soil depth
for both the toposequences (Table 7). The probable cause may be the shallow
plough layer beneath which the hard ploughpan existed. More than 99% roots
were confined to 0-10 cm soil layer. BR11 yielded the highest (4.5 t/ha), followed
by Biharibatraj, Magursail, and Jhingasail. The yield of local varieties ranged from
2.5-3.0 t/ha.
162

Table 6 :

Grain yield and ancillary characters of the early maturing breeding

lines/varieites grown under rainfed conditions at alimganj, T. aman.
1993.

Date of sowing : 4.7.93; Date of transplanting : High land (H) : 27.07.93;

Medium land (M) : 29.07.93; Low land (L) : 31.07.93

Experiment 5. Effect of Ploughpan management on the performance of rice under

rainfed lowland situation.
Materials and methods
The experiment was conducted at Alimganj during T. aman, 1993. Four
ploughpan manipulations were: i) control (undisturbed ploughpan or normal tillage,
i.e. farmers practice of 6-8cm depth), ii) reduced ploughpan through deep tillage
(12-15cm), iii) reduced ploughpan through deep tillage (18-20cm)and iv) ploughpan
perforation through dhaincha (Sesbania aculeata) rooting with normal tillage. Three
varieties, BR11, BR25 and Biharibatraj and two breeding lines-BR4974-45-9-2 and
BR1870-88-1-11 were tested. The experiment was laid out in split plot design with
3 replications having ploughpan manipulation in the main plot and varieties/lines
in the sub-plot. The unit plot size was 4 x 4m.
In the ploughpan perforation treatment, dhaincha seeds were sown on 23rd
April, 1993 at the rate of 60 kg/ha and 90 days old dhaincha plants were harvested
at ground level and removed from the experimental plot. Phosphorus, potassium
and sulphur were applied at 25:35:20 kg/ha as basal. Nitrogen was applied at the
rate of 80 kg/ha in 3 equal splits as top dress.

163

Table 7.

Root weight and distribution pattern at different soil profile depths at

Alimganj, Rajshahi, 1991 (mean of 10 spots for each site).
% Distribution of root

Root weight (mg/cm 3 )

Site

Soil depth (cm)

0-10

10-20

Soil depth (cm)

20-30

0-10

10-20

20-30

Low toposequence
Site 1

5.94

0.031

0.004

99.41

0.52

0.52

Site 2

6.85

0.037

0.002

99.43

0.54

0.03

Site 3

9.48

0.063

0.002

99.32

0.66

0.02

High toposequence
Site 1

6.20

0.030

0.002

99.49

0.48

0.03

Site 2

6.51

0.042

0.001

99.34

0.64

0.02

Site 3

8.70

0.046

0.003

99.44

0.53

0.03

Twenty-five-day old seedlings were transplanted with a spacing of 20 x 20 cm, on

July 30,1993. Cultural practices were done as and when necessary. Observations were taken
on plant height, tiller number, grain yield, and root mass density upto 30 cm soil depth
with 10 cm interval. Soil analysis was done by following the standard procedures.
Results and discussion
Ploughpan existed at 9 cm soil depth and continued upto 17 cm. The hard
pan gradually developed due to excessive clay (50%) in the soil, terraced land type,
puddling, for growing mono rice crop and shallow ploughing by country plough.
Reduced ploughpan through deep tillage and ploughpan perforation through
dhaincha root increased root mass over farmers practice in all the varieties and
breeding lines (Table 8). Reduced ploughpan through dhaincha rooting treatment
gave higher root mass at 10-20 cm soil depth in all the varieties/lines. More than
95% of root mass was confined to 0-10 cm soil layer.
Reduced ploughpan through deep tillage and ploughpan perforation through
dhaincha roots significantly increased the grain and straw yield over farmers
practices. The promising lines, BR4974-45-9-2 and BR1870-88-1-1 and BR11 produced
identical yields (Table 9). Soil penetration resistance, measured after harvesting at
different soil depths, was higher(2.0-2.25 kg/cm) in normal tillage plot than in the
ploughpan manipulated plots at 10-15 cm soil depths. Its values reduced (0.75-1.0
kg/cm) in ploughpan manipulated plots through dhaincha rooting at 10-15 cm soil
depth. (Table 10).

164

Table 8.

(mg/cm 3 )

Effect of plouhgpan management on the root mass density*

at flowering stage. Alimganj T. Aman, 1993.
Varieties**

Ploughpan management
V1

V2

V3

V4

V5

Soil depth 0-10 cm

4.04

4.20

3.28

3.94

3.18

(97)

(98)

(98)

(98)

(98)

Reduced ploughpan through (1215 cm)

4.09

4.25

3.32

3.88

3.19

deep tillage

(96)

(96)

(97)

(96)

(96)

Reduced ploughpan through (18-20 cm)

3.38

4.16

3.32

3.95

3.03

deep tillage

(95)

(95)

(97)

(95)

(95)

Ploughpan perforation through dhaincha

4.10

4.32

3.45

4.05

3.26

rooting with farmer practice

(95)

(96)

(96)

(96)

(96)

Farmer practice (6-8 cm) deep tillage

Soil depth 10-20 cm

0.08

0.09

0.05

0.08

0.07

(3)

(2)

(2)

(2)

(2)

0.16

0.18

0.10

0.15

0.14

(4)

(4)

(3)

(4)

(4)

0.21

0.23

0.12

0.21

0.16

(5)

(5)

(3)

(5)

(5)

Ploughpan perforation through dhaincha

0.23

0.22

0.14

0.14

0.15

rooting with farmer practice

(50)

(5)

(4)

(4)

(4)

Farmer pratice (6-8 cm) deep tillage

Reduced ploughpan through (1225 cm)

deep tillage
Reduced ploughpan through (18-20 cm)
deep tillage

Figures in parenthesis represents percent distribution.

**

V1 = BR 497445-9-2 ; V2 = BR 1870-88-1-1; V3 = BR 11; V 4 = Biharibatraj;

V5 = BR 25

165

Table 9.

Effect of Ploughpan management on the yield (t/ha) of two promising

lines and three rice varieties under drought prone environment.
Alimganj, T. aman, 1993.

Ploughman management

Farmer practice (6-8 cm) deep tillage

Reduced ploughpan through (12-15 cm)
deep tillage
Reduced ploughpan through (18-20 cm)
deep tillage
Ploughpan perforation through
dhaincha rooting with farmer practice
Mean
LSD (0.05)
Farmer practice (6-8 cm) deep tillage
Reduced ploughpan through (12-20 cm)
deep tillage
Reduced ploughpan through (18-20 cm)
deep tillage
Ploughpan perforation through
dhaincha rooting with farmer practice
Mean
LSD (0.05)
*

V1

V2

V3

V4

Grain Yield (t/ha)

5.1
2.4
3.0
5.3

V5

Mean

2.9
3.3

3.9
4.3

4.6
4.9

4.7
5.0

5.2

5.2

5.3

3.1

3.5

4.5

5.3

5.2

5.6

2.7

3.5

4.5

5.0
0.63

5.0

5.3

2.8

3.3

5.9
6.1

5.4
5.8

6.1

5.9

6.0

6.1

5.8

6.1
0.25

5.7

Straw Yield (t/ha)

5.1
5.8
5.2
5.7

4.5
3.9

5.3
5.5

5.9

4.8

5.7

5.7

5.5

5.0

5.6

5.8

5.4

4.8

V 1 = BR 4974-45-9-2 ; V2 = BR 1870-88-1-1; V3 = BR 11; V4 = Biharibatraj; V5 = BR 25

Table 10.

Soil penetration resistance (Kg/cm 2 ) as affected

management at different depths. Alimganj, 1993

Ploughpan Management

by

ploughpan

Soil depth (cm)

5

10

15

20

25

30

0.5

2.0

2.2

2.2

1.5

1.0

Reduced ploughpan through (12-15 cm) 0.5

deep tillage

1.5

1.5

1.0

0.1

0.1

Reduced ploughpan through (18-20) cm 0.5

deep tillage

1.0

1.5

1.0

0.5

0.7

Ploughpan perforation through

dhaincha rooting.

0.7

1.0

1.2

0.5

0.5

Farmer practice (6-8) deep tillage

166

0.5

REFERENCES
BBS (1992) Bangladesh Bureau of Statistics, Dhaka, Bangladesh.
BRRI (Bangladesh Rice Research Institute) (1992) regional station, Rajshahi research report for 1991 wet
season, Rainfed lowland rice research consortium site at Rajshahi, Bangladesh.
Haque M E, Chang T T, Tepora N M, G C Loresto (1988) Root and shoot characters of rice ( Oryza sativa
L.) in relation to drought resistance aeroponic and hydroponic culture. Bangladesh J. Plant Breeding and
Genetics 1 (l&2): 18- 25.
IRRI (International Rice Research Institute (1992) Challenges and opportunities in less favourable
ecosystem: Rainfed lowland rice. IRRI information series no. 1, IRRI, Los Banos, Philippines.
OToole J C, Chang T T (1978) Drought and rice improvement in perspective. IRRI Research Paper Series
No. 14.:IRRI, Los Banos, Philippines.
Manalo E B (1975) Agro-climatic survey of Bangladesh. The Bangladesh Rice Research Institute and the
International Rice Research Institute.
Russell M B (1959) Water and its relation to soils and crop. Advances in Agron. 11 :1-122.

167

MODELLING THE SOIL-PLANT-WATER BALANCE OF

RAINFED LOWLAND RICE
M.C.S. Wopereis1, M.J. Kropff 2 and T.P. Tuong2

SUMMARY
A soil-water balance module (PADDY) and a module predicting the response
of rice to drought (DSTRESS) were developed and linked with a crop growth
simulation model (ORYZA1) to simulate rice growth and production in rainfed
lowland rice ecosystems. The combined PADDY-DSTRESS-ORYZA1 model
(ORYZA-W) was validated using two field experiments in the Philippines. Measured
and simulated changes in ponded water depth under flooded soil conditions were
in good agreement. In one of the field experiments, temporary drought was induced
at different stages. The model satisfactorily predicted change in root zone water
content, leaf area index, total aboveground dry matter, and panicle dry weight across
drought treatments over time. The model was subsequently used to predict rainfed
rice yield as a function of soil hydraulic properties and long-term weather data (25
yr) in Tarlac Province of Philippines. Risk involved in growing rainfed rice was
quantified by calculating yield probability distribution for seven major soil types
under rice cropping. Coupling ORYZA-W to a GIS allowed a spatial analysis of risk
in Tarlac Province.

INTRODUCTION
Process-based simulation models can be used as a tool to unravel some of the
complexity and variability of rainfed lowland rice ecosystems. Such models allow
detail analysis of experimental data, or extrapolation of research findings to other
environments.
Quantifying the responses of rice to drought stress is essential for predicting
the impact of soil and weather conditions on rice production. For lowland rice, grown
in puddled soils, there is little information on the link between soil-water status and
crop response, although drought is generally seen as a major cause of yield loss in
rainfed rice production system. Existing rice growth simulation models use standard
drought stress responses often derived for other crops.
The man-made puddled layer in lowland rice soils is often effective reducing
water loss through percolation to deeper soil layers. The effect of puddling on the

1 Department of Theoretical Production Ecology, Wageningen Agricultural University, P.O. Box 430, 6700 AK
Wageningen, The Netherlands
2 International Rice Research Institute, P.O. Box 933, 1099 Manila, Philippines.

hydraulic conductivity of the various layers is, however, not well understood. Drying
of previously submerged rice soils creates cracks that may extend through the plow
sole at the bottom of the puddled layer. This can cause a drastic and often irreversible
increase in water losses due to increased percolation rates. Existing soil-water balance
modules do not consider such changes and are not directly applicable to puddled
rainfed lowland rice soils.
A new soil-water balance module, PADDY, and a drought stress module,
DSTRESS, were developed and coupled to the model ORYZAI, (Kropff et al., 1993;
IRRI, 1993) for use in rainfed ecosystems. The combined ORYZAI-DSTRESS-PADDY
simulation model (ORYZA-W) was validated using two experiments conducted at
IRRI in 1991 and 1992, and used to predict rainfed rice yield on a regional scale for
a province in the Philippines.

MATERIALS AND METHODS

The crop simulation model ORYZA1 is documented elsewhere (Kropff et al.,
1993). Input requirements for ORYZA1 are :
*

Geographical latitude ;

Daily data on solar radiation and minimum and maximum temperature ;

Plant density ;

Date of crop emergence or transplanting ; and

Parameters
describing
the
morphological
characteristics of the rice variety.

and

physiological

In water-limited environments, water availability, light, and temperature

determine the crop growth rate (provided no nutrient limitations occur). Wopereis
(1993) derived functions that describe a number of physiological and morphological
responses of the rice crop to soil-water content of the root zone. These functions
were incorporated in the module DSTRESS. The module PADDY was developed to
predict the soil-water status of a puddled root zone. Both modules are described in
detail below. The linkage between ORYZAI, PADDY, and DSTRESS is illustrated in
Figure 1. PADDY and DSTRESS are written in FORTRAN and make use of the
Fortran Simulation Environment (van Kraalingen, 1991). A detailed users manual
for ORYZA-W is in preparation.
Description of PADDY
An irrigated puddled soil profile consists of a muddy layer with little
resistance to water flow, a less permeable plow sole and, the non puddled subsoil.
The muddy layer gradually increases in bulk density with depth but is treated here
as a uniform layer. In PADDY, it is assumed that the first two layers of the soil profile
comprise the muddy layer and the plow sole, respectively. Time step of integration
in PADDY is 1.

169

Fig. 1 :

Linkage between the modules ORYZA1, PADDY, and DSTRESS

In PADDY, precolation rate can either be an input or can be calculated using

an iterative Newton-Raphson Procedure (Wolfram, 1991) comparing fluxes through
the plow sole and nonpuddled subsoil. Capillary rise is calculated based on
integration of the Darcy equation for steady, upward, vertical flow (Gardner, 1958).
Capillary rise can be ignored if the groundwater table is sufficiently deep, or if roots
have not reached the nonpuddled subsoil (see for details Tuong et al., 1993). The
potential rates of evaporation from soil and water and crop transpiration with soil
background (dry field) or water background (wet field) are derived from the Penman
reference evapotranspiration calculated from daily weather data (vapor pressure,
temperature, wind speed) using a Penman type equation (Penman, 1948). Actual
transpiration rate is estimated from the potential transpiration ratg and a drought
stress factor calculated by DSTRESS, described below. Actual evaporation rate from
soil is calculated by assuming that cumulative evaporation is proportional to the
square root of time (Stroosnijder, 1982). The rate on the first day is assumed to be
60% of potential soil evaporation. In reality, puddled clay qoil will probably dry out
faster because of the rapid appearance of shrinkage cracks, which may increase the
evaporation surface to more than double in a puddled clay soil (Fujioka and Sato
,1986). This aspect has not been taken into account in PADDY.
170

The continuous drying of a puddled clay soil results in shrinkage cracks and
subsidence of the soil surface. To simulate cracking of the puddled root zone,
knowledge of the soils shrinkage characteristic is needed (Fig. 2), where moisture
ratio v is defined as the volume of water Vw over the volume of the solid phase Vs,
and void ratio e is defined as the volume of pores Vp over the volume of solid phase.
The shrinkage characteristic is used to calculate the volume of pores per volume of
soil (e, m3 /m3 ), the volume of water per volume of soil (q, m3 /m3 ), the subsidence
of the puddled soil surface, and the change in crack volume (Bronswijk, 1988). It is
assumed in PADDY that shrinkage is irreversible and that the puddled muddy
topsoil gradually regains structure, a process that usually is referred to as soil
ripening. Total porosity e therefore, declines upon drying and will not increase if
the water content of the root zone increases. In PADDY, cracks are assumed to have
penetrated through the plow sole if its simulated moisture ratio drops below 1.2,
which for IRRI soil is equivalent to a soil pressure potential h of - 100 kPa (IRRI,
1992).

Fig. 2 .

Soil shrinkage characteristic of a puddled clay soil. Values in diagram indicate soil pressure
heads (cm).

If soil cracks have not yet reached the plow sole, it is assumed that all incoming
water is used to replenish the first soil compartment but that cracks will not close.
During this phase the percolation rate is zero. As soon as the water content of the
top compartment has reached saturation, water starts ponding again and the
percolation rate will be governed by the hydraulic conductivity of the plow sole.
The amount of water that can be stored in the top compartment is calculated taking
into account the changes in volume and porosity of the top compartment due to
cracking. If cracks are deep enough to reach the plow sole (i.e.h<100 kPa), all water
in excess of field capacity will be drained from the top compartment. Because of the
171

soil ripening process, the conventional field capacity concept (i.e. volumetric water
content at h=-10 kPa) is hard to use for puddled soil conditions. The field capacity
water content of the topsoil was, therefore, defined as 95% of total porosity. For the
nonpuddled subsoil, the conventional definition for field capacity was used. In
PADDY, water that drains from the cracked root zone will fill up soil layers below
the root zone up to field capacity. Any excess water will be drained at a maximum
rate equal to the saturated hydraulic conductivity of the subsoil horizon. The water
content of soil compartments below the groundwater table depth is reset to
saturation.
The required data for PADDY are :
*

Saturated soil-water content of puddled topsoil;

Saturated soil-water content of ripened topsoil and nonpuddled subsoil;

Soil-water content at wilting point of the nonpuddled subsoil (h= -1.5

MPa);

Soil-water content at field capacity of the nonpuddled subsoil (h=-10 kPa);

Saturated hydraulic conductivity of the plow sole, or infiltration rate;

Hydraulic conductivity characteristics of the subsoil (if groundwater table

is relatively deep, or if roots cannot reach the subsoil because of a plow
pan, capillary rise can be ignored and only the saturated hydraulic
conductivity of the subsoil is needed);

Groundwater table depth throughout the growing season (usually average

groundwater table depths will be used);

Soil shrinkage characteristics of puddled topsoil and nonpuddled subsoil

(optional; if not available, default values will be used).

Description of DSTRESS
DSTRESS is largely based on data from a greenhouse experiment conducted
at IRRI in the dry season (December - May) of 1992. In this experiment, the response
of lowland rice cultivars IR20 and IR72 grown in puddled clay soil to temporary
drought at different growth stages (transplanting, 2 wk after transplanting, midtillering, panicle initiation, and flowering) was studied. Morphological responses
(inhibition of leaf production, leaf rolling, appearance of dead leaves) and
physiological responses of the crop (reduction in transpiration rate and decrease in
development rate) were expressed as a function of soil moisture ratio of the root
zone. A similar approach was taken by Sinclair (1986) and McCree and Fernandez
(1989). During the experiment, the degree of leaf rolling was visually examined daily
at midday using a leaf rolling scale (1-5). A leaf score of 1 indicates the fipst sign of
leaf rolling, whereas a score of 5 means that the leaf has completely rolled up (after
O, Toole and Cruz, 1980). The experiment was repeated in the 1992 wet season. For
reasons of brevity, only results from the drought at mid-tillering treatment conducted
in the dry season are reported here (Fig. 3).
172

Fig. 3 .

Relationships between soil moisture ratio (cm 3 /cm-3 ) and (a) relative leaf growth, (b) leaf score,
(c) % dead leaves, and (d) relative transpiration rate in a greenhouse experiment conducted
at IRRl in the dry season of 1992 (from Wopereis, 1993).

Relative leaf growth (defined as the ratio between the leaf growth of stressed
plants and that of well-watered plants) decreased rapidly from 1 to 0 if soil moisture
ratios dropped below 1.7, i.e. leaf expansion of plants subjected to drought stopped
(Fig. 3a). Leaf rolling started at lower moisture contents, and leaf rolling score
increased from 1 to 5 within a relatively narrow range of soil moisture ratios (Fig.
3b). As drought progressed, the percentage of dead leaves increased rapidly as well
(Fig. 3c). Both leaf rolling score and percentage of dead leaves were linearly related
to soil moisture ratio. Transpiration rate per unit of area of plants subjected to drought
(Td) remained equal to that of well-watered plants (Tw), even if soil-water status
dropped nearly 50%. As soil-water content declined further, a decrease in relative
transpiration rate (defined as Td/Tw) was observed (Fig. 3d). Well-watered plants
and plants that were temporarily stressed in the vegetative phase did not differ
173

significantly in yield for either cultivar. However, flowering and maturity were
strongly delayed. Severe drought in the reproductive phase greatly reduced yields.
The following morphological and physiological plant responses to drought were
quantified for the different growth stages : a) inhibition of new leaf production, b)
leaf rolling, c) leaf senescence, d) decrease in relative transpiration, and e) decrease
in development rate in the vegetative stage. Responses a, b, and c followed each
other more or less sequentially, whereas response d) started at roughly the same
soil-water content as response a) and declined to zero when leaves were dead (end
of c) (Fig.3).
The effect of drought stress on development rate in the vegetative phase (e)
could not be directly measured and is therefore not shown in Figure 3. In DSTRESS,
the development rate in the vegetative phase used in ORYZA1 (Kropff et al., 1993)
is multiplied by a factor that increases from 0 to 1 between transplanting and
flowering. This means that the closer the development stage is to flowering, the
smaller the postponement effect. No delay in growth is simulated if drought occurs
in the reproductive phase.
In the greenhouse study, production of new leaves was strongly inhibited than
dative transpiration per unit leaf area during drought periods in the vegetative phase.
This means that CO2 assimilation continues, but the C produced cannot be used for
leaf production. In DSTRESS, excess C is stored in a pool and released for leaf production
as soon as drought stress is released. The 1992 dry season experiment was repeated in
the 1992 wet season.

Model validation
The ORYZA-W model was tested for flooded soil conditions using data from
a field experiment conducted at IRRI in the 1991 dry season (cv IR72) and described
in detail by Wopereis et al. (1994) and Bouman et al. (1994). Input variables were
rainfall, irrigation, evapotranspiration rate from daily weighing of pots installed in
the field, and groundwater table depths measured using piezometers. Average and
upper and lower extreme values for measured hydraulic conductivity of the plow
sole and the non puddled subsoil were used. Simulated and observed changes in
ponded water depth were compared.
For nonflooded soil conditions, the ORYZA-W model was tested using data
from a second experiment (field experiment 2) conducted in the 1992 dry season on
a 2,000 m2 field (cv IR72). Four drought treatments in four replications were tested
in a randomized complete block design in 10 x 5 m subplots. These subplots were
separated by bunds and hydraulically isolated by plastic sheets placed 0.6 m into
the soil. Prior to the start of the experiment, all plots were submerged for 10 d, then
plowed three times and harrowed three times using a water buffalo.
Drought was initiated at transplanting (T1), at mid-tillering (T2), at panicle
initiation (T3), or during the grain filling stage (T4) by simply draining the ponded
water from the plots. For comparision, a well-watered treatment was included (T0).
Two drought durations (D) were imposed based on the 1-5 leaf rolling scale of
O'Toole and Cruz (1980). For Dl, plots were rewatered to allow plant recovery from
174

drought when leaves showed initial leaf rolling (leaf score = 1). For D2, plots were
rewatered when leaves showed clear sign of leaf rolling (leaf score = 3).
All simulations were conducted using parameters obtained from the 1992 dry
season experiment with IR72 reported in Kropff et al. (1993), except for the
development rates of the vegetative and reproductive phase and leaf N concentration
as a function of time, which were derived from the well-watered TO plots.
Model application
The ORYZA-W model was used to predict rainfed rice yield for the wet season
(June-November) in the Province of Tarlac which is located in the northern part of
the Philippines on the island of Luzon (Fig. 4). It covers an area of approximately
300, 000 ha and comprises 17 municipalities with a total population of about 740,
000 (BSWM, 1992).
A soil map of Tarlac Province (1:50, 000), which was provided by the Bureau
of Soils and Water Management, Quezon City, Philippines, was digitized using the
geographic information system (GIS). The total number of mapping units was
reduced from 67 to 14 through generalization taking into account similarity in soil
properties and importance of the unit in terms of surface area. A similar approach
was taken by Bregt et al. (1989).
As a first qualitative step, soil unsuitable for rice growth, which included the
mountainous area and light-textured soils, were eliminated from the analysis.
However, light textured soils classified on the soil map as severely flooded because
of their proximity to a river were not excluded; it was assumed that rice grown on
such soils does not suffer from drought stress. Simulations were conducted for
potentially suitable soils only. Potential (irrigated) rice yield was simulated using
the model ORYZA1 for 25 yr of weather data derived from a meteorological station
in the center of the province.
Water-limited (rainfed) rice yield for the same set of 25 yr was simulated using
the ORYZA-W model. Guided by the detailed soil map representative profiles for major
soil type under rice cropping were located. The soil hydraulic properties needed in
ORYZA-W were determined for all soil horizons in each representative profile.
Procedures and results of these measurements are presented elsewhere (Wopereis et al.,
1993). Crop parameters for rice cultivar IR72 were derived from a wet-season experiment
conducted at IRRI in 1991 (Kropff et al., 1993). Simulations started at transplanting,
assuming that seedlings were 30 days old. Initial leaf area index (LAI), temperature sum
(sum of thermal units over a base temperature), and development stage of the seedlings
were taken from field experiment 2. Initial rooting depth was assumed to be 0.05 m.
After discussion with an expert from the Bureau of Soils and Water Management,
Quezon City, Philippines (W. Sanidad, pers. commun.), transplanting of rice was
assumed to start when cumulative rainfall exceeded 75 mm during seven consecutive
days after 1 June. Thickness of the puddled topsoil was set to 0.15 m with the plow sole
occurring between 0.15 and 0.20 m depth. At transplanting, the puddled topsoil was
assumed to be saturated with an initial ponded water depth of 0.05 m. Subsoil horizons
were assumed to be at field capacity (h= -10kPa).
175

Fig. 4.

Location of the Province of Tarlac in the Philippines.

The impact of groundwater table depth and thoroughness of puddling on

rainfed rice yield was investigated for each of the major soil types under rice
cropping. Groundwater table depth was varied between 0.5 and 1.0m. Wopereis et
al. (1992) determined the saturated hydraulic conductivity of the least permeable
176

layer (i.e. plow sole) in the top 0.2 m of a puddled clay soil at the experimental farm
of the International Rice Research Institute. Average value was 0.036 cm/d, with
95% confidence limits at 0.027 and 0.045 cm/d. In this study, two classes of puddling
(poorly puddled and well puddled) were considered and expressed in terms of the
hydraulic conductivity ks of the plow sole: well puddled, Ks (plow sole) = 0.01 cm/d
and poorly puddled Ks (plow sole) = 0.10 cm/d. Combined with the two water table
depths, four simulation series were created:
1 :

Ks plow sole=0.01 cm/d; water table depth = 0.5m

2 :

Ks plow sole=0.01 cm/d; water table depth = 1.0m

3 :

Ks plow sole=0.10 cm/d; water table depth = 0.5m

4 :

Ks plow sole=0.10 cm/d; water table depth = 1.0m

Probability distributions of rainfed rice yield for all soil types under rice
cropping were estimated. Maps of simulated rainfed rice yield at different levels of
cumulative probability were produced using the GIS software.

RESULTS AND DISCUSSION

ORYZA-W validation
The water balance module PADDY accurately predicted the changes in ponded
water depth for field experiment 1 (Fig. 5). Wopereis et al. (1993) tested the more

Fig. 5.

Simulated and observed changes in ponded water depth in field experiment 1 using soil-water
balance module PADDY.

177

Fig. 6.

178

Simulated (lines) and observed (symbols) total dry matter weights (t ha-1) for cv. IR72 in field
experiment 2 for all drought treatments : A is drought at transplanting, early recovery; B at
transplanting, late recovery; C at mid-tillering, early recovery, D at mid-tiliering, late recovery;
E at panicle initiation, early recovery; F at panicle initiation, late recovery; G at flowering, early
recovery; H at flowering, late recovery.

complicated differential SAWAH soil-water balance module (ten Berge et al., 1992)
using the same field data. Results from this study showed that the iteration procedure
used in PADDY to calculate the flux through the soil profile under flooded soil conditions
was as effective as the small time step calculations used in SAWAH. The observed and
simulated total above ground matter (Fig. 5), LAI, panicle dry weight, and root zone
water content (Fig. 6) were compared. The results indicated that ORYZA-W could
satisfactorily explain the differences in biomass production and soil water content across
drought treatments, although for root zone, water content predictions were less good
at severe stress and for drought at flowering.
Tarlac simulations
Potential yields in Tarlac Province varied from 5.4 to 6.7 t/ha (Fig.7), which
are considerably higher than the irrigated rice yields (2.5-3.5 t/ha) reported by BSWM
(1992). This discrepancy may be due to a number of factors, e.g. lack of fertilizer,
incidence of pest and diseases, etc. that wen' taken into account by the ORYZA-W
model.

Fig. 7.

Simulated (lines) and observed (symbols) soil water content m 3 /m -3 for cv. IR72 in field
experiment 2 for drought at transplanting (0-5 cm, late recovery, closed circles), drought at
mid-tillering (0-10 cm, late recovery, squares), drought at panicle initiation (0-10 cm, late
recovery, triangles), and drought at flowering (0-1 0 cm, no recovery, open circles).

Rainfed rice yields ranged from 0 to 6.7 t/ha. For reasons of brevity, the
variability of rainfed rice yield over 25 years for the four simulation series is shown
in Fig. 8 for two distinct soil types only : Zaragoza clay loam (light texture) and
Padapada clay soil series (heavy texture). Potential yields are also shown for
comparison.
179

Fig. 8.

180

Potential and rainfed rice yield calculated for 25 consecutive wet seasons and four simulation
scenarios on Zaragoza and Padapada soil.

Fig. 9.

Cumulative distribution functions for rainfed rice yield on Zaragoza and Padapada soil.

181

Comparison of rainfed rice yield with potential yields quantifies the yield gap
between fully irrigated and rainfed production. This information indicates the yield
loss farmers will experience due to lack of irrigation water, under otherwise optimal
growing conditions. Production risk was quantified by calculating cumulative
probability functions for rainfed rice yield for each soil type (Fig. 9). For Zaragoza,
a shallow groundwater table had a positive effect on grain yield due to increased
capillary rise to the root zone. For Padapada, this effect was almost non- existent.
Poor puddling resulted in yield losses for both soils, especially for Zaragoza.
The hydraulic conductivity of the plow sole was an important determinant
of rainfed rice yield for light-textured soils with a relatively permeable subsoil, like
the Zaragoza soil series. If no information on this soil parameter is available, a
constant percolation rate determined for the various soil types may be used as an
input for the PADDY soil-water balance module (Wopereis et al., 1994; Bouman et
al., 1994). Tuong et al. (1994) showed that percolation losses toward and into bunds,
and the effect of poorly puddled sites may be important in areas with a relatively
permeable subsoil. More complex numerical models that allow for lateral flow into
the bunds (eg. Walker and Rushton, 1984) are needed under these circumstances.
On a regional scale, one-dimensional models, such as the PADDY soil-water balance
module can still be used, provided a constant percolation rate is assumed,
incorporating both vertical and lateral percolation losses.
Simulated rainfed rice yield was mapped at the 10 and 90% cumulative
probability levels using the GIS software. Simulations were conducted assuming
average ks values of the plow sole determined in the laboratory and a water table
depth of 1.0 m. The Zaragoza soil series occupies a large part of the potential
rice-growing area in Tarlac Province (Wopereis, 1993). Growing rice under rainfed
conditions in that province is, therefore, risky.
The approach outlined above can only result in a broad overview of yield
losses due to drought in Tarlac. Soil types were characterized using measurements
conducted at one representative site only. Spatial variability of soil hydraulic
properties or thickness of soil horizons was not taken into account.

CONCLUSION
The GIS and crop simulation modelling can be used to quantify rice yield
losses due to drought at a regional level. Rainfall variability has a strong impact on
yield variability in Tarlac Province. Field experiments conducted for 1 or 2 yr in such
environments may give misleding results. Long-term weather data are needed to
determine probability distributions of crop yield to perform an economic evaluation
(Anderson, 1991). Unfortunately, there is a lack of long-term weather data in many
rice-growing countries in Asia as was also the case for the study presented here
(detailed weather data were only available for one station in the entire province).
Supplementary irrigation increased wet-season rainfed rice yields and
reduced yield variability. Irrigation may also increase the potential for a dry-season
crop (eg. mungbean), which would boost total production and income per year
relative to rainfed conditions. The scope for a dry-season crop after rice could be
182

investigated using the PADDY soil-water balance and good explanatory model for
the dry-season crop.
FUTURE RESEARCH NEEDS
One of the major problems in the application of crop-soil models is the lack
of data on soil hydraulic functions. Data bases, that relate these functions to soil
characteristics derived from soil survey data, are needed. Maintenance and
installation of weather stations should be promoted partly through research
consortia. It is important to have a few well-selected and well-maintained sites that
can be carefully monitored. Only if good data sets are available can simulation
models be used to extrapolate new technologies or to identify constraints or
opportunities in rice production. More research is needed on how to deal with limited
data in crop modelling studies.
The rainfed rice model ORYZA-W can be used to investigate yield losses due
to drought in rainfed, puddled environments and may also be used to quantify the
benefits of improved irrigation facilities. ORYZA-W is also a starting point for
simulation studies on rice upland crop rotations. The model, however, needs further
validation for a broader range of environmental conditions and for different rice
varieties. PADDY needs to be expanded to include simulation of water table depth.
Sensitivity analyses to investigate the importance of variability in model-input
parameters must be conducted.
The increased water-use efficiency of rice-based cropping systems is becoming
increasingly important. This can be done by improving irrigation facilities,
introducing water-saving techniques, and adjusting the planting time and/or
cropping system to maximize rainfall utilization. For any of these approaches, a
thorough systems analysis to evaluate the different solutions for different
environments is needed. One of the most promising water-saving techniques is dry
seeding of rice. The drought stress functions used in this study were all derived for
puddled soil conditions. More research is therefore needed on drought stress
responses of dry seeded rice and their relation to root zone water content. Special
attention should be paid to root development and the effect of drought on root
growth.

REFERENCES
Anderson J R (1991) A framework for examining the impacts of climatic variability. Pages 3-17, In : Climatic
risk in crop production: Models and management for the semiarid tropics and subtropics. R.C. Muchow
and J.A. Bellamy, eds. CAR International, Wallinglord, UK.
ten Berge H F M, Jansen D M, Rappoldt K, Stol W (1992) The soil water balance module SAWAH : Users
guide and outline. CABO-TPE Simulation Report Series, no. 22 CABO Wageningen, The Netherlands, 78
pp. + appendices.

183

Bouman B A M, Wopereis M C S, Kropff M J, ten Berge H F M, Tuong T P (1994) Understanding the

water efficiency of flooded rice fields. II. Percolation and seepage losses. Agricultural water management
(in press).
Bregt A K, Wosten J H M, Denneboom J (1989) GIS: belangrijk hulpmiddel bij bodemfysische interpretatie
van bodemkaarten. Landinrichting 29 (3): 9-14 (in Duch).
Bronswijk J J B (1988) Modelling of water balance, cracking and subsidence of clay soils. J. Hydrol. 97 :
199-212.
BSWM (1992) Soils/land resources evaluation project Tarlac Province. Physical environment,
agro-socio-economic and land evaluation. Department of Agriculture, Bureau of soils and Water
Management, Quezon City, Philippines.
Fujioka Y, Sato K (1986). On the drying of clayey paddy field (11). Tarns. JSIDRE, 26: 1-7. (in Japanese
with english abstract).
Gardner W R (1958) Some steady state solution of the unsaturated moisture flow equation with application
to evaporation from a water-table. Soil Sci. 85: 228-232.
IRRI (1992) Programme report for 1991. International Rice Research Institute, P.O. Box 933, 1099 Manila,
Philipppines.
IRRI (1993) Programme report for 1991. International Rice Research Institute, P.O. Box 933, 1099 Manila,
Philipppines.
Kraalingen D W g van (1991) The FSE system for crop simulation. Simulation Report CABO-IT no. 23,
Centre for Agrobiological Research, P.O. Box 14, Wageningen The Netherlands, 77 pp.
Kropff M J, van Laar H H, ten Berge H F F (1993) ORYZA1, a basic model for irrigated rice production.
International Rice Research Institute, P.O. Box 933,1099 Manila, Philippines, 78 pp.
McCree K J, Fernandez C J (1989) Simulation model for studying physiological water stress responses of
whole plants. Crop Sci. 29: 353-360.
O'Toole J C, Cruz R T (1980) Response of leaf water potential, stomata1 resistance and leaf rolling to water
stress. Plant Physiology. 65:428-432
Penman H L (1948) Natural evaporation from open water, bare soil and grass. Proceeding of the Royal Sciety
of London Series A 193: 120-146.
Sinclair T R (1986) Water and nitrogen limitations in soybean grain production. I. Model development.
Field Crops Res. 15: 125-141.
Stroosnijder L (1982) Simulation of the water balance. Pages 175-193. In: Penning de Vries, F.W.T. and van
Laar, H.H (Eds). Simulation of plant growth and crop production. Simualtion Monograps. Pudoc,
Wageningen.
Tuong T P, Wopereis M C S, Singh A K, Kropff M J, ten Berge, H F M (1993). Modeling lowland rice
agrohydrology. Presented at the Dec. 14-17 1993 ASAE International Winter Meeting. Paper No. 932528.
Tuong T P, Wopereis M C S, Marquez J, Kropff M J (1994) Mechanisms and control of percolation losses
in puddled rice fields. Soils Science Society of American Journal. (In press).
Walker S H, Rushton K R (1984) Verification of lateral percolation from irrigated rice fields by a numerical
model. J. Hydrol. 71: 335-351.

184

Wolfram S (1991) Mathematica. A system for doing mathematics by computer Second edition.
Addison-Wesley Publishing Company Inc., Redwood City, California, 961 pp.
Wopereis M C S, Wosten J H M, Bouma J, Woodhead T (1992) Hydraulic resistenace in puddled rice soils
measurement and effects on water movements Soil & Tillage Research 24: 199-209.
Wopereis M C S (1993) Quantifying the impact of soil and climate variability on rainfed rice production.
Ph.D. Thesis, Wageningen Agricultural University, The Netherlands. ISBN 90- 5485-147-3.
Wopereis M C S, Kropff M J, Hunt E D, Sanidad W, Rouma J (1993) Case study on regional application
of crop growth simulation models: Tarlac Province, Philippines. Pages 27-46. In : Bouman, B.A.M., van
Laar, H.H. and Zhaoqian, W. (Eds). Agro- ecology of rice-based cropping systems. ISBN 90- 73384-18-4.
Wopereis M C S, Bouman B A M, Kropff M J, ten Berge H F M, Maligaya A R (1994) Understanding the
water use efficiency of flooded rice field,s I. Validation of the soil-water balance model SAWAH.
Agricultural Water Management (in press).

185

DROUGHT RESPONSE OF DRY SEEDED RICE AND ITS

MODELLING
A.K Singh 1 , T.P. Tuong 1 , K.T. Ingram 2 , R Pernito1 and J, Siopongco 1

SUMMARY
Dry seeded rice (DSR) provides an option to increase crop production and
intensity by utilizing early-season rainfall, escaping late-season drought and
reducing turnaround time for a post-rice crop. Experiments conducted in 1992 and
1993 at the Tarlac Rainfed Lowland Consortium Key-site evaluated and quantified
the response of DSR to drought in a rainfed lowland environment. The treatments
were different dates of seeding and varying water regimes, each having three levels.
The seeding dates were May 18, June 11, and July 8 in 1992, and May 27, June 11,
and August 4 in 1993. The three water regimes were (i) totally rainfed conditions,
(ii) initially irrigated conditions to ensure complete emergence and then rainfed, and
(iii) fully irrigated conditions. The ORYZA1 model was parameterized for DSR
through vegetative and reproductive development rates and tested to simulate DSR
growth under fully irrigated conditions. Transpiration and evaporation were
determined by daily weighing of 20x20x40cm lysimeters installed in the fields. The
final yield was not significantly affected by the various water regimes except for the
first seeding date. Under rainfed conditions, the second seeding date resulted in the
highest yield. Rice plants transpired at a potential rate as soil water content of the
mot zone decreased from saturation to field capacity (-0.03 MPa matric potential).
Below field capacity, the relative transpiration rate decreased linearly with decreasing
soil moisture. This linear decrease appeared to be independent of the fertility level
and climatic conditions. ORYZA1 satisfactorily estimated dry matter production and
yield under fully irrigated conditions.

INTRODUCTION
About 27% of the total rice area in the world has been classified as rainfed
lowland, which contributes less than 18% to global rice production. More than 95%
of the world's total rainfed lowland area is in Asia (IRRT, 1993). Even a small increase
in the productivity of these regions would add significantly to global rice production.
With the availability of modern, short-duration, high yielding varieties, dry seeded
rice (DSR) provides an option to increase the productivity in rainfed lowland
ecosystems. DSR can be sown even before the onset of the wet season, which permits

1 International Rice Research Institute. P.O. Box. 933, Manila, Philippines.

2 Peanut CRSP Management Office, University of Georgia, Griffin,

Georgia 30223-1797, U.S.A.

utilization of early-season rainfall (Tuong et al., 1993). Seeds of DSR remain viable
in the soil for as long as 40 days without detrimental effect on germination (Furoc
et al., 1978). The early harvest of DSR enables the crop to escape and/or minimize
the adverse effects of drought at the end of the wet season, and leaves enough time,
as well as available soil moisture and late-season rainfall, for a second crop (Saleh
et al., 1993). Since DSR is done in nonpuddled conditions, it saves on turnaround
time for post-rice crops.
The uncertainty of rainfall may expose the crop to drought of varying intensity
and duration during the growing season. This is one of the key factors limiting yields
of rainfed rice. Advancing the seeding date in DSR aggravates this problem because
the crop is established before the steady monsoon season rain. DSR, because of its
very nature of cultivation, is associated with many problems starting with crop
establishment, high weed competition and periodic dry spells. Since the soil is not
puddled, DSR represents a totally different environment for rice root development
as well as for water and nutrient uptake compared with transplanted rice (TPR).
There is a dearth of precise and systematic information on the soil-water-plantatmospheric interactions in relation to DSR. It is, therefore, essential to quantify the
effects of drought on soil-water extraction and plant performance in DSR grown
under rainfed lowland ecosystems. It will enable development and/or modification
of a model linking environmental conditions to dynamics of soil water and plant
growth which could be utilized for identifying conditions and locations suited for
DSR, the risks associated with it and the year to year yield variations.
The main objectives of this paper are :
1)

to quantify the effect of drought on the transpiration rate of dry seeded rice; and

2)

to test the ORYZA1 model, a validated crop growth model for potential
production of TPR (Kropff et al., 1993), for DSR under fully irrigated
conditions, as a first step toward a full scale modelling of DSR under rainfed
conditions.

MATERIALS AND METHODS

Experimental site
The experiments were conducted at the Tarlac Rainfed Lowland Consortium
Key site located at Village Masalasa, Victoria, Central Luzon, Philippines. The annual
rainfall of Victoria is 1610 mm, with 1310 mm being recorded from June to October
(16 year average). The wet season is characterized by frequent typhoons. The general
texture of the soil profile is clay loam averaging 38% clay, 42% silt, and 20% sand.
The surface soil layer (0-10 cm) had 0.85% organic carbon content. The soil water
characteristic (pF) curve is shown in Figure. 1.
Performance of rainfed DSR
TO evaluate the performanceof rainfed DSR, field experiments were conducted
during the wet season (WS) of 1992 and 1993 in Tarlac. Since the amount and
187

Fig. 1.

Water retention characteristics of the surface soil (0-20 cm) at Tarlac.

frequency of rainfall are highly variable and uncertain, the seeding dates were varied
during the same growing season in order to expose the crop to drought at various
stages under field conditions. The treatments were three seeding dates and three
water regimes. The dates of seeding (Dl, D2, and D3) were May 18, June 11, and
July 8 in 1992 and May 27, June 11 and August 4 respectively in 1993. Water regimes
were (i) totally rainfed (IO), (ii) irrigated initially to ensure complete emergence and
then rainfed (11), and (iii) fully irrigated control (I2). All plots were separated from
surrounding plots by plastic sheets embedded 60 cm deep around each plot. The
experiment was laid out in a split plot design with water regimes as the main plot
and seeding date as subplot treatment, replicated five times. The subplot was 7.5x11.5
m. Nitrogen (120 kg/ha) was applied in four toydressed equal splits, i.e. 15 days
after emergence (DAE), maximum tillering, panicle initiation, and heading. Twenty
five kg P/ha and 45 kg K/ha were also applied. Seeds were sown at a rate of 100 kg/ha
in rows 20 cm apart and 2 cm deep. IR72 was used in both experiments. The experimental
plots were kept free of weeds and pests, following recommended practices.
Emergence count of the crop under IO was monitored daily in four 1-m-long
rows in each subplot from seeding until a constant number, i.e. final emergence, was
attained. Plant samplings from two 0.5m-long-rows in each subplot were carried out
for biomass production, leaf area index (LAI) and leaf N content. Samplings were
188

done at 15 DAE, mid-tillering, maximum tillering, panicle initiation, between panicle

initiation and flowering, and maturity.
Rainfall, wind speed, solar radiation, relative humidity, and maximum and
minimum temperatures were also recorded daily at a weather station located
adjacent to the experimental fields.
Measurement of transpiration
Two experiments were conducted during the 1993 dry season (DS) and 1993
wet season (WS) to investigate the effect of drought on DSR, by weighing lysimeters
(20x20 cm cross-sectional area and 40 cm deep, with an undisturbed soil profile of
0-30 cm) installed in the fields. The soil surface within the lysimeter was covered
with a plastic sheet to prevent evaporation, allowing for water loss through
transpiration only. Transpiration was derived from changes in the daily weights of
the lysimeters which were also used to calculate the average soil water content in
the lysimeter.
The 1993 DS experiment consisted of two main treatments each having two
levels. The main treatments were water regimes, stressed (described below) and fully
irrigated, and fertilizer application rates of 60 and 120 kg N/ha, replicated four times.
The 1993 WS experiment consisted of three seeding dates described earlier and two
water regimes, stressed (described below) and fully irrigated, replicated three times.
Stress was imposed in the lysimeters by withholding irrigation at the panicle
initiation stage (1993 DS experiment) and at the maximum tillering stage (1993 WS
experiment). In the wet season, the lysimeters were also sheltered from rains during
the stressed period to prevent alleviation of drought by rain. Stress was relieved by
supplying water when the plants registered a leaf-rolling score of 5 (OToole and
Cruz, 1980). The rice variety used was IR72.
The effect of water stress on the plants was expressed in terms of the relative
transpiration rate, defined as the ratio of the actual transpiration (Ta) and potential
transpiration (Tp) rates. Imposition of stress slowed down plant growth and leaf
area development resulting in reduced leaf area in stressed plants compared with
well-watered plants. The potential transpiration recorded for the well-watered plants
(Tpns) would be higher than the potential transpiration of the plants having a leaf
area equivalent to that of stressed plants. Tpns was therefore corrected to the potential
transpiration rate (Tp) of a plant having the same leaf area as the stressed plant on
a daily basis (Wopereis, 1993). The steps involved are explained below.

where LAIs is the leaf area index of stressed plant and LAIns is the leaf area
index of fully irrigated plants.
Since destructive sampling could not be carried out from the lysimeters until
maturity, plants were sampled from an equivalent area within the plots of fully
189

irrigated treatments to obtain LAIns at the start and at termination of stress. LAIns
and LAIs were considered equal at the initiation of stress. After the stress was
relieved, the stressed plants began transpiring at a potential rate (Tp) that was less
than the transpiration rate of well-watered plants (Tpns). The difference between Tp
and Tpns was due to the difference in the LAI of stressed and non-stressed plants
at the end of the stress period. LAIs at the end of the stress period was derived from
LAIns Tp; and Tp ns using the relation.
where ln is natural logarithm.
Daily values of LAIns and LAIs were worked out from the LAIs values obtained
at the initiation and end of stress, assuming that the change in LAI during that period
was a linear function of time. Daily CF values were then derived using equation 2.

Modelling of DSR
The ORYZAI crop growth model was tested for simulation of growth and
yield of fully irrigated DSR. ORYZAI was originally developed and validated to
simulate growth and yield of TPR rice under potential production conditions (Kropff
et al., 1993). The model requires radiation, temperature, planting dates, crop
development, and relative growth rates as well as time course of leaf N content as
input. Measured LAI can also be input; otherwise, it is predicted by the model.
Development rate parameters and leaf N content are important for simulation
of rice growth for a given environment (Kropff et al., 1993). However, the
development rate of DSR will differ from that of TPR. As a first step, therefore, the
development rate parameters of DSR were derived from the first sowing date of I2
in the 1993 WS experiment described earlier, based on seeding, flowering, and
physiological maturity dates using the program DRATES (Kropff et al., 1993).
Measured leaf N content was used as input. Simulations predicted the total biomass
and panicle dry weight of I2 treatments for all sowing dates in 1992 and 1993. LAI
was also simulated, but the results are not presented in this paper.

RESULTS AND DISCUSSION

Rainfall pattern
Figure 2 presents the cumulative rainfall from seeding to harvest at different
seeding dates. It varied from slightly above 1000 mm in D3-92 WS to more than 1700
mm recorded in D2-93 WS, thereby providing a wide range of water availability
during those two years. Varying the seeding dates, therefore, generated different
water availability conditions during the same growing season.
Emergence of rainfed DSR
In the rainfed treatment (I0), emergence of DSR was influenced by the rainfall
after sowing and the soil water content at the time of seeding. At the first seeding,
the soil water content in the top layer was around 10-11% (Table 1), which is much
190

Fig. 2.

Cumulative rainfall during 1992 and 1993 in dry seeded rice (DRS) as a function of days after
sowing.

below the water held by the soil at wilting point (-1.50 MPa matric potential). In this
treatment, emergence was delayed by almost a week and the emergence percentage
in 1993 was 64% (Fig. 3). The emergence in 10 of the first seeding of the 1993 WS
was triggered by 69 mm of rain (8 days after seeding) and the emergence count
stabilized at a cumulative rainfall of less than 100 mm. Germination in the subsequent
seeding dates was 80% or more because of higher initial soil-water content and
greater amount of rainfall.
Table 1.

Soil water content in 0-10 cm layer at sowing time in 1992 and 1993.

Sowing date
1992

1993

Soil water content

(cm3 /cm3 )

18 May
11 June
8 July

0.11
0.16
0.20

27 May
11 June
4 Aug.

0.10
0.21
0.23

191

Fig. 3.

Effect of sowing date on emergence of rainfed (I0) DSR IR72. Numbers in parentheses are
cumulative rainfall totals (mm) at which emergence count became constant.

Grain yield
The final yield was not significantly affected by the various water regimes
except on the first date of seeding (Table 2). Under rainfed conditions the second
date of seeding gave maximum yields because of better water availability throughout
the growing season. Heavy rainfall, coinciding with flowering of the crop slightly
reduced yield in I2 for the second sowing date in 1992 compared with other
treatments. Tuong et al. (1993) reported similar observations of yields remaining
unaffected even though drought spells delayed emergence and other plant
phenological stages under rainfed conditions.

Transpiration ratio
For Ta/Tp, the results of the experiments conducted during 1993 DS and WS
were pooled and are shown in Figure 4 as a function of soil water content. Ta/Tp
during the DS were for 60 and 120 kg N/ha, whereas that for WS was at 120 kg
N/ha. In well-watered plants during DS, transpiration rates were 10 mm/day and
reduced to 2 mm/day in the stressed plants at the time of drought termination (at
leaf rolling score of 5). During WS, transpiration rate of well- watered plants was
192

5-6 mm/day and was reduced to 1.5 mm/day or less at the later stages of stress.
For all seasons and N levels, the rice plant was able to transpire at a potential rate
until the soil moisture was around field capacity, i.e. at a soil pressure potential of
-0.03 MPa. Beyond this soil water content, Ta/Tp decreased linearly with decrease
in soil water content, (R2 =0.9254). In a study of transplanted rice under non
nitrogen-limiting conditions, Maligaya et al. (1993) and Wopereis (1993) obtained
similar trends in the relative transpiration rate as a functiol of soil moisture ratio,
which could be translated into soil water content. In their study, Ta/Tp began
declining at a soil moisture ratio of 1.7, representing soil water content below field
capacity when water stress was initiated at mid- tillering stage for both IR20 and
IR72. The critical soil moisture at which Ta/Tp starts to decline needs to be
investigated further as it may be both crop and stage-specific as well as being
influenced by the crop establishment method.
Table 2.

Effect of seeding date and water regimes on the yield of dry seeded rice
IR72.

Seeding date*
Dl

D2

D3

*
**

Water regime

Yield** (t/ ha)

1992

1993

I0

2.9b

3.7ab

I1

2.5b

3.4b

I2

4.0a

4.la

I0

3.8a

4.3a

I1

4.0a

3.8b

I2

3.3b

3.9ab

I0

3.8a

3.8a

I1

3.8a

3.7a

I2

4.0a

3.8a

Dl refers to 18 May in 1992 and 27 May in 1993; D2 refers to 11 June in both 1992 and 1993; D3
refers to 8 July in 1992 and 4 Aug. in 1993.
Means having a common letter in the same column are not significantly different at the 5% level by
DMRT.

Plants respond to the imposition of stress through stomatal closure and leaf
rolling. Stomatal closure reduces water loss through transpiration whereas leaf
rolling directly reduces the leaf area exposed to the environment, which also
contributes to reduced transpiration losses. Reduction in transpiration rate directly
affects the rate of photosynthesis. This relationship (Fig. 4) representing reduction
in relative transpiration rate as a function of soil water content can be used to link
a potential production simulation model to production under water-limiting
193

Fig. 4 .

Relative transpiration rate (Ta/Tp) of DSR as a function of soil water content (WS=wet season;
DS=dry season; N60=60kg Nha; N120=120 kg N/ ha).

environments through a water balance model. This could lead to a model that can
be used to quantify the effects of water stress on crop growth and yield of DSR.

Modelling of dry seeded rice

The crop development rates for the vegetative and reproductive phases of
DSR derived from the first sowing date of I2 in 1993 using DRATES were 0.000576
and 0.001568, respectively. The corresponding values for TPR IR72, as reported by
Kropff et al. (1993), were 0.000751 and 0.00168. The crop development rate of DSR
during the vegetative phase was markedly different from that of TPR. The simulated
results for total above ground dry matter production for all sowing dates
corresponding to I2 of 1992 and 1993 are shown in Figure 5. The dry matter prediction
is generally within 10% of the observed values except in the case of the third seeding
in 1992. As indicated by Figure 5, the biomass is overestimated for all seedings in
1992. This is also reflected in the panicle dry weight predictions (Fig. 6). One of the
possible causes for this overestimation could be the existence of field conditions
inadequate for potential production, caused by untimely typhoons and heavy rains,
leading to yield reductions. Despite these discrepancies, the fitness of simulated and
measured biomass and panicle dry weights was satisfactory (R2 = 0.8699).
The model needs to be evaluated further to simulate growth of DSR under
water-stressed conditions. It also requires linking up with a module that can simulate
the dynamics of soil water under rainfed lowland situations, where perched water
194

Fig. 5 .

Observed (0) and simulated () values of dry matter production for all sowing dates of I2.

195

Fig. 6 .

Comparison of observed and simulated dry matter production and panicle dry weight of
well-watered DSR IR72.

tables occur frequently and for several weeks an unsaturated layer can be observed
below a saturated one in the soil profile. Initial testing of process-based water balance
model SAWAH has indicated that it can be used successfully under these situations
(Tuong et al., 1993).
196

CONCLUSION
Results of our experiments indicate that Ta/Tp was a sound parameter for
characterizing stress as it appeared to be independent of N application rate and the
prevalent climatic conditions. ORYZA1 satisfactorily simulated the dry matter and
yield of DSR under fully irrigated conditions. Further work is required to modify
the model to simulate crop growth under water- stressed situations, and coupling
the crop component with a water balance model for rainfed lowland ecosystems.

REFERENCES
Furoc R E, Magbanua R D, Gines H C, Morris R A (1978) Identification of criteria for late dry-seeded rice
planting. Paper presented at the 9th Annual Scientific Meeting of the Crop Science Society of the
Philippines, 11-13 May, 1978, Iloilo City, Philippines.
IRRI (1993) International Rice Research Institute. IRRI rice almanac, P.O. Box. 933, Manila, Philippines.
Kropff M J, Van Laar H H, ten Berge H F M (1993) ORYZA1-A basic model tor irrigated lowland rice
production : Simulation and Systems Analysis for Rice Production (SARP). International Rice Research
Institute, P.O. Box 933, Manila, Philippines.
Maligaya A R, Wopereis M C S, Kropff M J (1993) Drought stress responses of two lowland rice cultivars.
Paper presented at the 9th Federation of the Crop Science Society of the Philippines Scientific Meeting, 10-14
May, 1993, Aklan, Philippines.
O'Toole J C, Cruz R T (1980) Response of leaf water potential, stomatal resistance and leaf rolling to water
stress. Plant Physiology 65: 428-432
Saleh A F M, Lantican M A, Bhuiyan S I, Agua M M (1993) Increasing productivity of droughtprone rainfed
lowlands in Northern Luzon, Philippines. Paper presented at the Symposium of the rainfed Lowland Rice
Research Consortium, 8-13 Feb. 1993, Semarang, Indonesia.
Tuong T P, Ingram KT, Siopongco J D, Confessor R B, Boling A A Singh U, Wopereis M C S (1993) Performance
of dry seeded rainfed lowland rice in response to agrohydrology and N-fertilizer management. Paper
presented at the Symposium of the Rainfed Lowland Rice Research Consortium, 8-13 Feb. 1993 Semarang,
Indonesia.
Wopereis M C S (1993) Quantifying the impact of soil and climate variability on rainfed rice production. Ph
D thesis, Agricultural University, Wageningen, The Netherlands.

197

RICE ROOT SYSTEMS DEVELOPMENT UNDER

RAINFED CONDITIONS
G. Pantuwan1, K. T. Ingram 2 and P.K. Sharma3

SUMMARY
Field experiments to quantify root growth of rice cultivars under rainfed
conditions were conducted at the Ubon Rice Research Centre, Northeast Thailand,
during the 1991 and 1992 wet seasons. Root length, root mass, and root:shoot ratio
were measured at panicle initiation and anthesis. Genotype, soil hydrology, and soil
type significantly affected root length density (RLD) and root mass density (RMD).
The RLD and RMD of genotypes at plots subjected to alternate wetting and drying
were generally higher than those at plots that were continuously flooded. Greater
RLDs and RMDs were observed in loamy sand soil than in clay soil. Root:shoot ratio
was higher in loamy sand than in clay soil irrespective of hydrology. Roots in clay
soil exhibited less root branching than roots in loamy sand. Loamy sand soil subjected
to alternate wetting and drying may be an appropriate growing condition for
evaluating rice root growth.

INTRODUCTION
Optimum root growth is required for stabilizing shoot growth in plants
subjected to drought stress. A deep and dense root system has been shown to be a
heritable characteristic, and has been assumed to be essential for allowing plants to
avoid or tolerate drought. Thus, rice breeders have attempted to incorporate
favourable root characteristics into rice genotypes for drought- prone, rainfed
lowland areas.
A measurable variation in root system characteristics of rice genotypes has
long been recognized (Yoshida and Hasegawa, 1983; O'Toole and Bland, 1988). It
has also been shown that soil physics, hydrology, and agricultural practices influence
rice root growth (Maurya and Ghildyal, 1975; Sharma et al., 1987; Thangaraj et al.,
1990; Ogunremi 1991). Rice roots, therefore, exhibit a large degree of phenotypic
plasticity. Coupled with genotype x environment interactions, root phenotypic
plasticity makes evaluation of root system characteristics difficult. The soundness
of information on rice root traits depends on whether plants were grown in an
environment where full expression of their growth potential is realized. Such
growing environments need to be identified.
1 Ubon Rice Research Centre, P.O. Box 65, Ubon 34000, Thailand.
2 Peanut CRSP Management Office, University of Georgia, Griffin, Georgia, 30233-1797, U.S.A.
3 Department of Soil Science, H.P. Krishi Vishwavidyalaya. Palampur-176062, H.P.. India

This paper summarizes information from research on the growth and

development of rice root systems on various soil types and hydrological regimes in
rainfed areas of Northeast Thailand.
MATERIALS AND METHODS
Experiments were conducted during the 1991 and 1992 wet seasons at the
Ubon Rice Research Centre in Northeast Thailand, to quantify root system growth
and development under rainfed conditions. In 1991, three rice cultivars KDML105,
RD9, and IR46 were transplanted (28 d old seedlings) on puddled soils at three
locations giving different combinations of hydrology and soil texture :
1)

High toposequence,
nonflooded;

loamy sand,

alternate

wetting and

drying,

mostly

2)

Low toposequencc, loamy sand, flooded most of the time during the growing
season; and

3)

Low toposequence, clay soil.

Six rice cultivars IR-20, KDML 105, NSG 19, RD9, Chieng Saen, and IR46 were
used in the 1992 experiment. Seedlings (30 d old) were transplanted in puddled
loamy sand soil, at high and low toposequence position. Experiments in both years
were laid out in randomized complete block design with five replications. Fertilizer
application, in both years, consisted of 25-25-25 kg/ha of N, P, and K incorporated
before transplanting and 50 kg/ha of N topdressed at panicle initiation (PI).
The genotypes used in the experiment were chosen to represent contrasts in
stature, photoperiod sensitivity, and drought susceptibility. KDML 105, (photoperiod
sensitive) and NSG19 (photoperiod insensitive) are traditionally selected, tall
cultivars, resistant to drought stress. IR20 (short stature high yielding) and Chieng
Saen (tall traditional cultivar) are known to be susceptible to drought stress. Cultivars
IR46 and RD9 are photoperiod insensitive, medium-statured, improved, and high
yielding.
Genotypes were sampled for roots at the same physiological age at PI and
anthesis. Root and shoot sampling was done on median plants at randomly selected
locations within plots. Roots were collected from four hills in each replication using
a 20 x 20 x 50 cm metal monolith sampler. Root sampling was centered over a hill.
Each soil core was sectioned into five 10 cm layers. Roots were washed free of soil
over a metal screen (1 mm mesh size). Root length was measured with a Comair
root length scanner (Commonwealth Aircraft Corporation Limited, Australia). Roots
were oven-dried at 70 for 48 h to determine mass. Root length density (RLD) and
mot mass density (RMD) were calculated as the quotient of mot length and oven
dry weight, and the volume of the soil sample. Above ground biomass was
determined in a manner similar to root mass. Root : shoot ratios (root oven dry
weight/shoot oven dry weight) were calculated from the data.
In the 1991 experiment, root form and proliferation were observed using the
pinboard technique. Pins were driven through the sampler in one hill in each
199

replication at anthesis. The monolith was carefully dug out, and removed from the
ground. Roots in the sampler were washed free of soil, allowing observation of
relative root distribution in the soil profile.

RESULTS AND DISCUSSION

Root length and root mass
Root length density was significantly affected by soil type, hydrology, and
crop stages. The 1991 experiment showed that the RLD of the three rice cultivars
was higher at high toposequence position than at low toposequence position.
Genotypes exhibited differences in RLD at the high toposequence position only (Fig.
1). The 1992 experiment, using six rice cultivars, gave results similar to those of the
1991 experiment (Fig. 2). Differences in root length density among genotypes,
particularly at anthesis, were greater at high toposequence position. Alternate
wetting and drying at the high toposequence position might have led to increased
mot length density of the cultivars. Kawata and Ishihara (1959) found that in aerobic
soil, drying periods induced the development of root hairs and nodal roots.
Root length density was generally higher in loamy sand than in clay soil (Fig.
1). Observations of root proliferation, using pinboard technique, confirmed that there
was greater root branching in loamy sand than in clay soil (Fig. 3). Greater root
branching in loamy sand soil may be due to its greater porosity, being dominated
by pores of relatively large radii, thus providing room for root growth (Kar et al.,
1979).

Fig. 1.

200

Root length density at anthesis stage of 3 rice cultivars at 3 toposequence positions (Pantuwan
et al 1992, unpubl).

Fig. 2.

Root length density of 6 rice cultivars (a) KDML 105, (b) MSG 19, (c) RD 9, (d) Cheing Saen,
(e) IR 20 and (f) IR 46 growth at high and low toposequence positions (Pantuwan et al 1993,
unpubl. data).

Fig. 3.

Root systems of three rice cultivars grown in (a) clay loamy sand (b) low toposequence and
(c) loamy sand high tosposequence position.

201

Root mass density (RMD) exhibited the same trend as RLD. Roots in loamy
sand produced greater RMD than the roots in clay soil (Fig. 4). Root mass density
was higher at high toposequence than at low toposequence positions (Fig. 5). Higher
RMD in the high toposequence may be related to the higher RLD observed at this
toposequence position.

Fig. 4.

Effect of soil texture on root mass density of 3 rice cultivars sown at low toposequence (Adapted
from Sharma, 1994).

Fig. 5.

Root mass density of 6 rice cultivars (a) KDML 105, (b) MSG 19, (c) RD 9. (d) Chieng Saen,
(e) IR 20 and (f) IR 46 grown at high and low toposequence positions (Pantuwan et al, 1993
unpubl. data)

202

More than 80% root length and 90% root mass were obtained in the top 20
cm soil layer, irrespective of soil hydrology (Tables 1, 2). Both RLd and RMD
decreased exponentially with the soil depths in all genotypes. This trend was similar
to previousstudies on rice roots (Pantuwan et al., 1992; Sharma et al., 1987). Genotypes
on the low toposequence position had higher percentage of roots in the top l0cm
soil layer than genotypes on the high toposequence position. This relationship was
reversed in the 10-50 cm soil layer.
Table 7.

Percent root length distribution of six rice cultivars grown under two
toposequence positions at Ubon Rice Research Centre, Thailand.
(Pantuwan et al., 1993, unpubl. data).
Root length distribution(%) at soil depth

Cultivar

0.10
(cm)

10-20
(cm)

20-30
(cm)

30-40
(cm)

0-10
(cm)

10-20
(cm)

Panicle initiation
High toposequence
KDML 105
NSG19
RD9
Chieng Saen
IR20
IR46
Mean
Low toposequence
KDML105
NSG19
RD9
Chieng Saen
IR20
IR46
Mean

20-30
(cm)

30-40
(cm)

40-50
(cm)

Anthesis

71.43
2.01
60.61
2.60
65.09
3.08
70.01
1.63
63.50
1.78
72.60
2.24
67.2
1.97

17.11
1.62
20.40
1.60
18.50
3.06
16.55
1.48
19.92
0.73
13.88
1.99
17.7
0.99

8.97
0.73
16.09
1.52
11.72
1.36
11.91
1.06
13.58
0.86
10.33
1.98
12.1
1.02

2.483
0.251
2.894
0.364
4.691
1.306
1.533
0.365
3.024
0.670
3.195
1.336
2.970
0.421

61.63
5.01
58.69
3.45
62.63
1.75
67.71
3.88
63.50
3.81
65.00
3.35
63.2
1.25

27.48
5.17
28.00
2.65
25.06
2.15
21.55
3.96
28.82
3.13
19.47
2.30
24.6
1.38

9.46
0.69
12.15
0.88
10.97
1.09
9.69
0.95
9.62
11.23
13.80
2.29
10.95
0.71

1.410
0.318
0.983
0.275
1.233
0.260
0.892
0.225
1.013
0.201
1.636
0.484
1.195
0.117

0.017
0.008
0.176
0.042
0.100
0.026
0.152
0.049
0.050
0.011
0.094
0.044
0.098
0.024

71.58
5.26
58.79
1.89
66.01
2.34
67.79
3.09
66.96
1.94
71.70
1.51
67.1
0.75

18.21
3.31
28.27
1.65
25.70
1.99
25.71
2.22
26.67
2.02
19.04
2.74
23.9
0.48

7.73
1.42
11.57
1.11
7.46
1.36
5.59
1.36
5.85
1.58
8.42
1.66
7.8
0.24

2.475
1.010
1.376
0.689
0.827
0.372
0.896
0.297
0.522
0.166
0.836
0.268
1.155
0.10

77.34
3.11
76.85
2.57
79.96
2.12
70.17
1.88
76.20
2.97
79.84
2.51
76 76
0.91

17.28
2.38
16.99
3.76
15.43
2.22
21.67
1.99
17.91
2.28
16.69
1.91
17.66
0.84

3.98
0.88
3.87
1.00
2.66
0.45
5.82
0.88
4.46
0.94
2.13
0.59
3.82
0.43

0.879
0.208
1.854
1.185
1.185
0.195
1.856
0.173
1.107
0.168
0.750
0.156
1.284
0.14

0.323
0.035
0.430
0.091
0.685
0.191
0.477
0.061
0.318
0.049
0.586
0.069
0.470
0.027

203

Table 2.

Percent mot mass distribution of six rice cultivars grown under two
toposequence positions at Ubon Rice Research Centre, Thailand.
(Pantuwan et al., 1993, unpubl. data).
Root mass distribution(%) at soil depth

Cultivar

0-10
(cm)

10-20
(cm)

20-30
(cm)

30-40
(cm)

0-10
(cm)

10-20
(cm)

Panicle initiation
High toposequence
KDML 105
NSG19
RD9
Chieng Saen
IR20
IR46
Mean
Low toposequence
KDML105
NSG19
RD9
Chieng Saen
IR20
IR46
Mean

20-30
(cm)

30-40
(cm)

40-50
(cm)

Anthesis

75.66
2.53
75.75
3.21
75.24
1.77
83.63
0.90
76.91
1.84
78.36
1.08
77.6
1.29

15.37
21.65
13.52
21.48
15.09
2.27
11.06
0.53
14.83
0.89
11.27
1.56
13.5
0.79

8.03
20.96
10.36
1.85
8.73
1.58
5.08
0.67
7.74
1.13
9.90
1.58
8.3
0.77

0.941
0.302
0.371
0.093
0.935
0.358
0.233
0.042
0.525
0.099
0.480
0.055
0.581
0.l20

77.07
0.48
73.96
2.24
73.27
1.32
75.63
2.56
78.09
1.23
76.86
2.59
75.8
0.77

17.09
1.06
18.94
1.89
19.57
0.89
18.79
2.29
17.24
0.71
16.66
2.23
18.0
0.49

5.31
0.57
6.26
0.49
6.29
1.03
5.08
0.60
3.66
5.68
1.85
1.09
5.41
0.40

0.519
0.008
0.15 0.001
0.804
0.037
0.253 0.003
0.857
0.013
0.230 0.003
0.478
0.039
0.055 0.002
0.983
0.030
0.140 0.004
0.618
0.009
0.104 0.002
0.710
0.023
0.082 0.006

83.22
2.73
72.25
1.76
75.52
2.27
80.20
1.79
76.39
0.94
82.03
1.31
78.3
1.73

12.84
1.59
20.64
1.56
20.07
1.35
16.77
1.49
19.90
1.6
14.46
1.62
17.4
1.35

3.18
1.18
6.71
0.69
4.22
1.17
2.79
0.78
3.45
0.85
3.27
0.50
3.9
0.59

0.860
0.334
0.391
0.175
0.181
0.048
0.239
0.084
0.258
0.104
0.235
0.069
0.361
0.104

83.72
2.69
78.63
1.53
82.89
1.74
85.25
1.06
85.49
1.86
82.98
2.54
82.8
0.93

14.13
1.90
16.49
2.03
14.20
2.07
13.66
0.68
12.15
1.06
14.99
2.20
14.3
0.59

1.74
0.70
3.20
0.80
2.24
0.39
2.54
10.35
1.99
0.86
1.73
0.53
2.2
0.23

0.377
0.029
0.251 0.008
0.047
1.637
0.458 0.008
0.580
0.090
0.190 0.013
0.047
0.504
0.101 0.011
0.297
0.074
0.011
0.193
0.040
0.258
0.054 0.006
0.054
0.609
0.211 0.009

Coefficients of variation of RLD and RMD within and between genotypes

were small near the soil surface and increased with soil depth in all growing
conditions. However, data at high toposequence position tended to have smaller
coefficients of variation than the data at low toposequence position.

Root : shoot ratio

In comparison to soil hydrology, soil texture had a more dominant effect on
204

root:shoot ratio (Table 3). Root : shoot ratios were higher in loamy sand than in clay
soil. High nutrient content associated with a clay texture, coupled with assured water
supply, apparently led to the production of more shoots than roots. This may indicate
that under nonlimiting nutrient and water supply, deep or extensive root systems
may not be required. Root : shoot ratio did not vary as much among genotypes as
among growing environments (see SEM values in Table 3).
Table 3.

Root : shoot ratios of three rice cultivars grown under rainfed

conditions.
Toposequence/soil

Cultivar1
Physiological
age

High/
loamy
sand

Low/
loamy
sand

Low/
clay

type
High/
loamy
sand

1991

Low/
loamy
sand

MeanSEM

1992

Panicle initiation
KDML105

0.37

0.34

0.26

0.16

0.19

RD9

0.36

0.28

0.25

0.18

0.17

0.250.039

IR46

0.41

0.36

0.34

0.17

0.15

0.290.059

Mean

0.38

0.33

0.28

0.17

0.17

0.015

0.024

0.006

0.012

KDML105

0.14

0.13

0.09

0.09

0.10

0.110.012

RD9

0.15

0.14

0.11

0.12

0.13

0.130.008

IR46

0.14

0.13

0.08

0.09

0.12

Mean

0.14

0.13

0.09

0.10

0.12

SEM

0.003

0.003

0.008

0.010

0.009

SEM

0.0028

0.260.046

Anthesis

0.130.016

CONCLUSION
Under rainfed conditions, rice root systems were significantly affected by
genotype, soil water regime, and soil type. Differences in root systems among
genotypes were large in loamy sand soil subjected to alternate wetting and drying.
Differences in root systems decreased under flooded conditions on fine textured
soils. Root:shoot ratios were affected more by soil type than by soil hydrology.
Coarse-textured soils subjected to alternate wetting and drying may provide
appropriate growing environment for the evaluation of root systems of rainfed rice.
205

REFERENCES
Kawata S, Ishihara S (1959) Study on the root hairs in rice plant. Proc. Crop Sci. Soc. Jpn. 27: 341-348.
Kar S, Varade S B, Ghildyal B P (1979) Pore size distribution and root growth relations of rice in artificially
synthesized soils. Soil Sci. 128:364-368.
Maurya P R, Ghildyal B P (1975) Root distribution pattern of rice varieties evaluated under upland and
flooded soil conditions. (Italian summary) Riso 24: 239-244.
Ogunremi LT (1991) lnfluence of bulk density and moisture regime of a permeablesoil on the performance
of lowland rice, Oryza sativa L. Trop. Agric. 68: 129-134.
O'Toole J C, Bland W L (1988) Genotypic variation in crop plant root systems. Adv. Agron. 41: 91-145.
Sharma P K, De Datta S K, Redulla C A (1987) Root growth and yield response of rainfed lowland rice
to planting method. Exp. Agric. 23: 305-313.
Thangaraj M J, O'Toole J C, De Datta S K (1990) Root response to water stress in rainfed lowland rice.
Exp. Agric. 26: 287-296.
Yoshida S, Hasegawa S (1982) The rice root system : its development and function. Page 97-114. In :
Drought resistance in crops with emphasis on rice. International Rice Research Institute, Philippines.

206

SOCIAL CONSEQUENCES OF STRESSED

ENVIRONMENTS
Thelma R. Paris1 , Abha Singh2 and Mahabub Hossain1

SUMMARY
Socio-economic studies were conducted in Chandpur and Mungishpur
villages of eastern U.P., interfaced with the biophysical characterization of selected
rainfed environments. This paper examines the labour contributions by gender
differentiation and social and economic groups and the problems of women farmers
in agricultural productions in different lowland rice ecosystems. The paper provides
suggestions which can improve crop and animal productivity as well as improve
the welfare of the entire family, particularly the females, and children. The study
revealed that the sexual division of labour is related to the social position of
the household and access to non-farm income. The crop-production, livestock
raising and non-farm activities are interrelated. The farmers can bear more risks
in rice cultivation if they have access to income from non-farm sources and would
be receptive to input-intensive technologies.

INTRODUCTION
In recent years, greater attention is being given to the complex, diverse
and risk prone (CDR) rainfed regions where mostly the poor people are located.
India has 40 million hectares of rice growing areas. While Northern and Southern
India has benefitted from the modern rice varieties and irrigation, the vast rainfed
areas in eastern India has lagged behind. Eastern India which is comprised of
Assam, Bihar, West Bengal, Orissa and eastern parts of Uttar Pradesh and Madhya
Pradesh has the highest incidence of rural poverty in the country. Although it
is the largest rice growing area in India, accounting 60% of the country's rice
area (26.8 million ha), only a third of the total rice production comes from this
region. About half (480 million) of the country's population living in this region
is largely dependent on rice farming. Macro level analysis of rice growing ecosystems
showed that out of the 26.8 m ha rice areas, only 21.2% is irrigated, 47.7% is
under rainfed lowlands and the remaining 14% is under deep and very deep
water ecosystems. The rainfed lowlands ape drought prone and submergence
prone (Singh, 1992).

1 Social Sciences Division, IRRI, P.O. Box 933, Manila Philippines

2 N D University of Agriculture and Technology, Kumarganj, Faizabad, India

Narendra Deva University of Agriculture and Technology (NDUAT),

Kumarganj, Faizabad, Uttar Pradesh is one of the agricultural research centres
under the ICAR-IRRI Collaborative Rainfed Rice Project in eastern India which
is involved in conducting on-farm research to increase rice production while
improving the efficiency of resources use in rice farming. The multidisciplinary
team involved in the project emphasizes on specific environments and
socioeconomic circumstances of the farmers of the region and focuses the attention
on development of technology consistent with resource endowments and farmer's
needs. However, constraints to increasing rice productivity, particularly in stressed
and fragile environments can not be well understood without knowing how farming
households, their female members in particular, allocate scarce resources to achieve
and maintain food security. The Social Sciences Division of IRRI in collaboration
with the NDUAT conducted socioeconomic studies which will be interfaced with
the biophysical characterization of selected rainfed environments in Uttar Pradesh,
eastern India.
This paper a) specifically examines the labour contributions of men and
women by different social groups in different lowland rice ecosystems; b) identifies
problems which women farmers face in agricultural production and; c) recommends
strategies which can improve crop and animal productivity as well as increase
the welfare of the entire family, particularly, the female members and children.
We hypothesize that due to adverse physical and socioeconomic conditions,
poverty drives women to increase their time and labour allocation to rice cultivation.
The proximity of rice based villages to towns and cities and higher wages from
non-farm employment induce male migration either seasonally or permanently and
this increases women's participation in farming. Further more, we hypothesize that
social classes and gender are important variables which affect rice cultivation
in eastern Uttar Pradesh.

METHODOLOGY
We selected two typical lowland rainfed sites, in eastern Uttar Pradesh,
namely Chandpur and Mungishpur which are different in terms of proximity
to the city and physical endowments. Chandpur is closer (3 kms) to Faizabad
city while Mungishpur is 28 kms away. Mungishpur represents drought prone
area which is favourable rainfed during years of high rainfall while Chandpur
represents shallow and submergence prone area which is favourable rainfed during
the years of low rainfall. In these sites, on-farm experiments on component
technologies e.g. varietal trials with a farming systems perspective are being
conducted by the research staff of NDUAT, in collaboration with the International
Rice Research Institute under the ICAR-IRRI Collaborative Rainfed Rice Program
funded by the Ford Foundation and International Fund for Agricultural
Development (IFAD).
We interviewed the principal heads (male and female) of 94 households
in Chandpur and 78 households in Mungishpur from 1992- 1993 after which
we classified them by social status (higher, scheduled caste and backward class).
208

RESULTS AND DISCUSSION

Social structure of the population
In both villages, majority of the population belongs to the lower castes, 88%
in Chandpur and 92% in Mungishpur. However, most of the land is owned by the
upper castes.
Gender gap in education, training and extension
There is a wide gap in terms of education by caste and by sex of adult workers.
Figure 1 shows that illiteracy rate is high among the lower castes than the higher
castes at both villages. In Chandpur, respectivily 92% and 84% of the total female
respondents are illiterate. Similarly, in Mungishpur, 98% and 100% of the total female
respondents are illiterate. However, the gap is even wider between males and
females. The high illiteracy rate of women than men is common in eastern India.
There seems to be a relationship between the environment and female illiteracy rates,
Five populous states (Andhra Pradesh, Bihar, Madhya Pradesh, Rajasthan and Uttar
Pradesh) account for more than half of Indias illiterate females. These five states
contain 89% of Indias districts where the rural female literacy rate is below 5%; 83%
of those are with rates of 5- 9%, and with 67% of those between 10-14%. Illiteracy
is especially widespread among rural women of the scheduled castes and scheduled
tribes. Many parents do not send their daughters to school because their labour is

Fig. 1.

Illiteracy rates of principal male and female farmers in rainfed villages of Faizabad, Uttar
Pradesh, India.

209

needed for agricultural and household tasks as well as for domestic chores such as
cooking and looking after young siblings especially when the mother has to work
in the fields. Parents are willing to forego the labour of sons and even incur cash
expenses on their education in anticipation of huge dowery in future (earnings and
security for the family) since the son will remain with the family even after marriage
and contribute to household earnings (Bennett, 1989).
In terms of access to agricultural information, women obtain information
about crop cultivation from their husbands, neighbours and other women. Despite
the important roles they play in agriculture, scientists include only male farmers
when conducting on-farm trials and only recently have begun to include women
farmers. Women also perceived themselves as housewives rather than farmers.
Cropping systems
In both villages rice followed by wheat is grown mainly for home
consumption. Land preparation for rice is done in June, transplanting or
broadcasting in July, weeding in August, harvesting in October and threshing in
November. Wheat is sown in January, harvested in March and threshed in April. In
June, land is again prepared for rice. Some farmers mix wheat and mustard and
grow lentils, grams and peas after rice. Sugarcane is also grown in the upland areas
from February to May. The availability of tubewells (8 in Chandpur and 3 in
Mungishpur) enables few farmers in Chandpur to grow vegetables which are sold
in the market.
Rice area and yields
Fig. 2 shows the average rice area holding across villages and by
socioeconomic status. Landholding devoted to rice is higher in Mungishpur than in

Fig. 2.

210

Average rice area by social class in rainfed villages of Faizabad, Uttar Pradesh, India.

Chandpur. Because of the proximity of Chandpur to Faizabad, the

centre, farmers diversify their income by growing vegetables and raise
production. In Mungishpur, farmers have lesser options for income
because they are far from the markets, thus they put more efforts in
production.

nearest urban
cattle for milk
diversification
increasing rice

In general, higher castes have larger rice farms (0.97 ha in Chandpur and 1.37
ha in Mungishpur) while the lower castes have smaller areas (0.5 ha and less).
Despite the smaller holding size in Chandpur, the average rice yields are
higher (2.0 to 3.24 tons/ha) compared with Mungishpur (less than 2 tons/ha). The
farmers from the higher castes obtained higher yields per hectare than the farmers
from the lower castes (Fig. 3). Richer farmers not only make more profit because
they are "receptive" to new rice technologies but they have cash and capital to
purchase for "input intensive" system. On the other hand, small farmers are less
willing to take the risks of applying inputs as their lands are often located in less
favourable conditions.

Fig. 3.

Average rice yield by social class in rainfed villages of Faizabad, Uttar Pradesh, India

Constraints to increasing rice yields

The lower rice yields in Mungishpur are due to many reasons. Drought is a
common environmental constraint which influences rice yields particularly when it
occurs during the grain filling stage. The average rainfall in U.P. is around 1,000 mm
and total monsoon days are about 125 days in a normal year. However, rainfall is
erratic and varies within the year. The coefficient of variability of yearly rainfall in
U.P. ranges from 25 to 30%. The approximate probability of deficient rainfall
(deficiency in rainfall numerically equal to or greater than 25% of the normal) in
211

U.P. is once in four years. Percentage occurrence of droughts of moderate class and
worse in the Kharif season is in the range of 20-25% in eastern U.P.. In addition, long
breaks in rainy days is a common phenomenon in India even during normal years.
Such breaks are common in U.P. and they result in contingent and invisible droughts.
Due to the extreme variability in the climatic conditions, farmers, particularly
in the drought prone areas are not willing to invest more inputs such as fertilizer
and herbicides.
Furthermore, the caste system which creates
distribution of resources is prevalent in eastern U.P.

inequality in access and

Labour use in rice production

Tables 1 and 2 show the total labour use in rice production. The total labour
input is higher in Chandpur than in Mungishpur (188 vs 132 labour days per ha.).
The contribution of female labour in rice production is higher than men in both
villages (84% in Chandpur and 67% in Mungishpur). The participation of adult
female workers is higher in Chandpur because men are more engaged in non-farm
jobs than in Mungishpur. In both villages, the social category of households also
determines the extent of female labour participation in rice farming. Based on the
household surveys, women from the higher caste do not provide physical labour
regardless of the size of their landholdings. Informal interviews with women from
the higher castes who have small landholdings indicated their desire to work in their
own fields or supervise hired labours, but are constrained to do so due to social
taboos. However, they are involved in production activities within the confines of
their homesteads such as seed management, storage and preparation of dry cowdung

Fig. 4.

212

Total labour use in rice farming by gender and social class in Chandpur and Mungishpur,
Faizabad, Uttar Pradesh India.

for fuel. In contrast, female family members among the lower castes work on their
farms as well as in other farms as exchange or hired labourers. In Mungishpur,
exchange labour is still widely practiced as a strategy for the lack of cash to pay for
hired labourers. Fig.4 shows higher labour input of female adult workers in rice
cultivation both in higher and lower caste farming families.
Table 1.
Activity

Prepare land
Apply manure/
chemical
Irrigate
Broadcast
Transplant
Weed
Harvest
Thresh
Totaln=94

Labour days per hectare in Chandpur, (rainfed lowland) 1991-92.

Family
Adult Adult
Male Female

Both

3.23
3.63

0.00
2.29

3.23
5.92

1.52
3.30
2.05
1.55
5.62
1.04
21.94

0.27
2.41
7.29
13.42
10.83
17.16
53.67

% contribution

Table 2.

Hired

30

70

Adult
Adult
Male Female

Both

0.95
0.65

0.00
0.12

0.95
0.77

1.79
5.71
9.34
14.97
16.45
18.20
75.61

2.44
0.80
1.64
0.12
1.07
0.15
7.82

0.00
0.06
40.93
33.02
15.20
14.81
104.14

100

Both

4.18
4.28

0.00
2.41

4.18
6.69

2.44
0.86
42.57
33.14
16.27
14.96
111.96

3.96
4.10
3.69
1.67
6.69
1.19
29.76

0.27
2.47
48.22
46.44
26.03
31.97
157.81

4.23
6.57
51.91
48.11
3272
33.16
187.57

100

16

84

100

Hired

Exchange

Total

Adult Adult
Adult Adult
Adult Adult
Adult Adult
Male Female Both Male Female Both Male Female Both Male Female Both

Prepare land
Apply manure/
chemical
Irrigate
Broadcast
Transplant

6.66
3.27

0.00
2.52

2.59
2.97
3.60

1.06
1.99

Weed
Harvest
Thresh
Totat n=78

3.12
4.18
2.85

% contribution

93

Adult
Adult
Male Female

Labour days per hectare in Mungishpur, (rainfed lowland) 1991-92.

Family

Activity

Total

29.24

36

6.66
5.79

1.56
0.60

3.65
4.96
8.66 12.26
12.22 15.34
10.53 14.71

0.10
0.55
4.66
2.17
2.06

14.28 17.13
51.26 80.50

1.31
13.01

64

100

29

0.00
0.00

1.56
0.60

0.00
0.00

0.00
0.00

0.00
0.00

8.22
3.87

0.05 0.15
0.13 0.68
12.07 16.73
7.86 10.03

0.00
0.00
0.50

0.00
0.00
2.16
1.54

0.00
0.00
266
1.74

269
3.52
8.76
5.49

1.81
0.98

2.16 6.59
1.13 4.31
7.69 43.45

7.05 9.11
4.15 5.46
31.31 44.32

71

100

0.20
0.35
0.15
1.20

16

6.49

84

100

33

0.00
2.52

8.22
6.39

1.11 3.80
2.12
22.89
21.62
19.39

5.64
31.65
27.11
25.98

19.41 23.72
89.06 132.51

67

100

213

Division of labour in rice production

In both sites, the division of labour between male and female workers is very
distinct (Fig. 5): Land preparation is exclusively done by men; irrigation and
application of fertilizers are predominantly done by men but women also assist
particularly when FYM is used as organic fertilizer for crops. Transplanting, weeding,
harvesting and threshing activities are highly dominated by female adult workers
from the lower caste. The Indian case clearly fits Ester Boserup's hypothesis that
where the female labourers come from a caste/ethnic group lower in social status
than the families with lands, the women in the latter avoid farm work to maintain
status. Similarly in the Philippines, female members in the irrigated farm hire female
workers to substitute for them in transplanting and weeding task (Boserup, 1970).
In between crop cultivation, women collect animal fodder, cultivate pulses, oilseeds
and maize and also perform most of the household responsibilities. These findings
indicate the strong correlation between class and gender in rice cultivation and land
distribution.

Off-farm and non-farm employment

In Chandpur, the male adult workers from the higher castes showed higher
earnings from off and non-farm work than the male adult workers from the lower
castes. Male adults from the higher castes have higher paying non-farm jobs because
they are better educated. In contrast, male adult workers in Mungishpur earn less
than 50% of what the male adult workers in Chandyur earn from off and non-farm
employment (Fig.6). These findings indicate that the greater opportunities for
non-farm employment for those with better education and skills pull the men away
from farm work. The upper caste families in Chandpur are faced with labour shortage
since most of their male members go to the cities and their female members do not
participate in farming.
While men earn from non-farm employment, women have less opportunities
due to their lack of education, mobility, training and skills. The major occupations of
male adult members are: teacher, bank manager, labour in iron factory, servant,
construction worker, military cook, juice seller, worker in biscuit factory, tailor, casual
labour in the university, rickshaw puller, home guard, barber in the military, watchman
and driver. A casual labour receives Rs 15 to 20 per day. Because women have less access
to paid jobs, they earn supplemental income by gathering, drying and selling "Mahua"
fruit flowers and selling them at 6-10 Rs/kg. They earn approximately Rs 500-600
(1US$=Rs30) within 15-30 days in May when the lands are not cultivated. They share
50% of the output if the trees are owned by higher castes.
Agricultural wage rates are higher in Chandpur than in Mungishpur due to
its proximity to Faizabad, the nearest urban centre. There is also difference in wage
rates for male and female workers for different operations, a male worker with
bullock receives Rs 25 per day while female workers receive Rs 10 for transplanting
and weeding. Transplanting is either paid on an individual or a contract basis.
Harvesters and threshers are paid through share basis. Thus when the yields are
low, the share of the harvester/thresher is also low.
214

Fig. 5.

Labour participation by gender in different rice farming activities, Chandpur and Mungishpur,
Faizabad, War Pradesh, India.

215

Fig. 6.

Average annual off-farm and non-farm income of adult male workers by social class in
Faizabad, Uttar Pradesh, India.

Animal systems
In both villages, animals constitute an integral part of a mixed farming
systems. Large animals (bullock and cattle) provide draft power for land preparation,
manure for crops household fuel and milk for home consumption and income.
Ownership of livestock allows farmers to diversify against risks. In Chandpur, more
than 50% of the total households interviewed have cattle and less than 38% have
bullocks. Because of the proximity of this village to the market, farmers particularly
the backward castes (Yadavs), raise cattle for milk which provides them an additional
source of income. In contrast, more households in Mungishpur raise bullocks than
cattle (Fig. 7). Due to the lower percentage of households which own bullocks in
Chandpur, less family male labour, and more income to hire tractors, more farmers
in this village use tractors for land preparation. In Chandpur, 41%, 18% and 9% of
the backward, scheduled and higher caste households, respectively use tractors. In
contrast, in Mungishpur, 15%, 13% and 8% of the backward, scheduled and higher
castes, respectively, use tractors (Fig.8).
216

Fig. 7.

Percent of households with bullocks and cattles in rainfed villages of Faizabad, Uttar Pradesh,
India.

Fig. 8.

Percentage of households using tractors in rainfed villages of Faizabad, Uttar Pradesh, India.

217

At both sites, animals are fed through cut and carry method due to lack of
common grazing lands. Farmers use crop residues, grasses and weeds from rice
fields to feed their animals. In Mungishpur, farmers grow Chari and berseem
in small plots. During the summer and part of the Kharif season (MarchOctober), rice and wheat straw are fed to the animals. One animal requires about
6 kgs of straw and 6 kgs of green fodder per day. Green fodder is most scarce
from April/May to mid June and available from August to September. Thus, a
few farmers grow berseem from October to March while Chari is grown from
March to August/September at about 0.03 ha of land. In Chandpur, very few
farmers have sufficient homesteads or field plots for fodder production, thus a
few of them rent government owned land to grow berseem or buy foddgr from
other farmers. Female family members are mainly responsible in collecting
grasses and weeds from distant places. With limited grazing lands, women have
to spend half a day daily to collect grasses and weeds, particularly in Chandpur
where there are more milch animals which are predominantly cared by female
family members. Twice a day they will feed and milk the animals, also clean the
shed or area where they are tied and collect the dung to make into fuel cakes.
During the drought periods, the only source of fodder and fuelwood is distant
forests, thus, women walk farther away from their homesteads to collect animal
fodder and fuelwood. With increasing deforestation and the degradation of the
common land, the burden falls on women and female children, whose time and
energy in gathering firewood has increased substantially.
Social consequences of stressed environments
What emerges from this initial assessment is the need for agricultural scientists
to understand not only the consequences of stress environments on rice yields, but
also the social consequences of drought on people especially on women who carry
the burden of fulfilling their multiple roles as a farmer, farm manager, wage labourer,
housewife and mother. With low crop productivity due to the fragile environments,
migration of men to the cities in search of work will continue to increase. The number
of female de facto household heads will also increase and will take over the
management of the rice farms. Increasing the demand for female labour in rice
production, processing, animal care, collection of animal fodder and fuelwood may
lead to changes in cooking habits (fewer meals), less time devoted to child care and
breast feeding. During periods of drought, the problem of securing food for the family
and fodder for the animals are hardships shouldered by women.
The goal of increasing rice productivity will be difficult to achieve if
management, particularly under stressed environments which require new
knowledge and skills, will be left to female family members, the majority of whom
are illiterate and without access to training and extension services. We also expect
that their work burden will increase as they face competing demands on their time
between households duties (caring for their children, collecting water, collecting
fuelwood, etc.) and agricultural responsibilities (management of crops, animals,
collecting animal fodder, processing, etc.) unless their male counterparts or children
share in that activities. The greater time spent outside the homes may have
218

detrimental effects on the welfare of the children unless women receive income
which they can use to provide the basic needs of their children.
The preliminary results of this study reveal that the environmental stress
which cause low crop productivity and income pushes men to engage more of their
time and attention in non-farm jobs and consequently, womens work in rice
cultivation increases. The low crop productivity will further exacerbate the disparity
between the higher and lower castes and women will further be burdened. Women
should be given access to education, training and extension services to enhance their
role in decision-making and in increasing farm productivity. They should be given
knowledge and skills on the timeliness of weeding and transplanting operations
which affect rice yields. Since they are the preservers of seeds, their indigenous
knowledge and varietal preferences are important in rice breeding programmes.
CONCLUSION
a)

The sexual divison of labor is related to the social position of the household
and access to non-farm income.

b)

Crop production, livestock raising and non-farm activities are interrelated.

c)

Farmers can bear more risks in rice cultivation if they have access to income
from non-farm sources and would be receptive to input-intensive
technologies.

Suggestions for research and extension

a)

Provide female workers access to education, training and extension programs :

Non-involvement of female workers in training and extension programs and
the illiteracy of female workers may be important constraints to the adoption
of improved farming practices.

b)

Include women farmers in evaluating new rice varieties : In both villages, improved
rice varieties such as NDR-423, NDR- 402, NDR-4001, Mahsuri which are
recommended for these rainfed environments have been distributed and
planted and are currently being evaluated by women cooperators.

c)

Introduce input-intensive cultivars : In the village where farmers have access to

non-farm employment and income.

d)

Introduce new species of animal fodder : Seedlings of new species of forage have
been introduced to increase the local supply of animal feeds and should be
evaluated by male and female farmers.

e)

Strengthen research-extension-farmers linkage : There are new initiatives in ICAR

whereby research-extension-farmer linkage is being strengthened among
agricultural universities and other institutions dealing with on-farm
experiments, for example, the Project Directorate for Cropping/Faming
System .

219

REFERENCES
Bennett Lynn (1989) Gender and poverty in Asia: Issues and Opportunities Concerning Women in the
Indian Economy. The World Bank, Washington, D.C. USA
Boserup Ester (1970) Womans role in economic development. St. Martins Press, New York, U.S.A..
Singh V P (1992) Rainfed lowland rice-based farming system in eastern India. Paper presented at the
Rainfed Lowland Rice Farming Systems Meeting, Yangoon, Myanmar.

220

DIFFERENTIAL GROWTH IN RICE PRODUCTION IN

EASTERN INDIA : AGROECOLOGICAL AND
SOCIO-ECONOMIC CONSTRAINTS
Mahabub Hossain and Alice Laboerte1

SUMMARY
Over the last two decades rice production in Eastern India grew at a
satisfactory rate of about 2.7 percent per year. But the growth was uneven across
states and districts particularly in Bihar, Orissa and Assam and Eastern Madhya
Pradesh where rice production failed to keep up with the population and income
induced growth in demand. Production performance was variable even within
homogeneous agroecological zones. Performance has been particularly poor in the
subhumid plateau regions of Chhotanagpur and Chhatisgarh, which also
experienced high rates of rural-urban migration.
Water control and the level of fertilizer use, which are highly correlated,
were important factors behind the spatial variation in productivity growth. The
level and the growth in rice yield was found to be negatively related with the
amount of rainfall, which suggests poor drainage and waterlogging as a
significant constraint to growth in rice production. Land tenure is a special social
problem as districts with higher incidence of tenancy and landlessness had lower
rice yields and slower growth. An inverse relationship is found between the size
of farm and the level and growth in rice yield, suggesting that the small size of
farm in the region is not a serious constraint to the adoption of input-intensive
technology. The size of farm, the incidence of irrigation and the instability in rice
yield are found important factors behind the spatial variation in fertilizer use
and the growth in rice yield.

INTRODUCTION
In the eastern part of India, rice is the principal staple food and agriculture is
dominated by the rice-based cropping systems. This geographical area comprises of
Assam and the small Northeastern States, West Bengal, Bihar, Orisa and the eastern part
of Madhya Pradesh (MT) and Uttar Pradesh (UP) and is inhabited by nearly 500 million
people. The region accounts for two-third of total rice area in India, but because of lower
productivity, contributes only half to the Indian rice production. Unlike other parts of
India where rice is grown under irrigated conditions, rice farmers of the region are

1 International Rice Research Institute. P.O. Box. 933, Manila, Philippines

dependent mainly on monsoon rains which is highly unpredictable, Until recently

the growth in rice production was sluggish, and the rice consumers were dependent
partly on imports from other parts (mainly Punjab and Haryana) where rice is grown
as a commenzial crop (Kumar and Rosegrant, 1993). Since the early 1980s, however,
growth in rice production is accelerating, and in the Ganges plain growth in output
has surpassed population growth.
In India, poverty and food insecurity are widespread in the eastern region.
According to recent estimates the poverty ratio is nearly 60% in Bihar, Orissa and
West Bengal, compared to less than 20% in Punjab, Haryana and Gujrat and less
than 30% in Tamil Nadu, Andhra Pradesh, Karnataka and Kerala (Dev et al., 1992).
The daily calorie intake is less than 2000 per capita in Bihar and Orissa, compared
to 2700 in Punjab and between 2400 to 2600 in Haryana, Rajasthan, Gujrat, Kerala
and Andhra Pradesh. The high incident of poverty can be partly related to the poor
performance of the rice sector which is not only the principal food but also an
important source of employment and income for rural population.
Since there is large unmet demand for food, per capita rice consumption is
expected to increase with economic prosperity and alleviation of poverty. As the
population of the region is still growing at 2.1% per year, the demand for rice may
grow by about 3.0% per year, if India could manage a 5% per year growth in gross
domestic product for this region. Sustaining a long-term growth of rice production
at 3.0% per year, in the face of unfavorable production environment, is a real
challenge to rice researchers and development practitioners for this region.
The main objectives of this paper are (a) to estimate the spatial variation in
the performance of the rice sector, (b) to understand the role of agroecological factors
in spatial variation in the level and growth of rice yields and (c) to assess the
interaction between agroecological and socioeconomic factors in constraining the
growth in rice productivity.
The study is based on historical data at the district level available from
Agricultural Situation in India (GOI, 1993) and Fertilizer Statistics (Fertilizer Association
of India, 1991). The data on agroecological and some socioeconomic characteristics
of the districts are taken from Huke and Huke (1982,1990). Growth rates in rice and
cereal production have been estimated from the time series data using
semi-logarithmic trend lines. To assess the role of agroclimatic and biophysical factors
in explaining the differential performance, we have classified the districts into
homogenous agroecological regions as defined in India (Sehgal, et al., 1990) and
compared the production performance across the zones. We have then applied a
multivariate regression analysis on the district level data to estimate the separate
impact of selected agroecological and socioeconomic factors. The paper concludes
with a summary of major finding and implications for rice research.
REGIONAL VARIATIONS IN PRODUCTION PERFORMANCE
In Eastern India rice production increased at 2.7 percent per year over the last
two decades (Table 1). The overall performance was satisfactory considering that
population grew at 2.2 percent per year during this period. The per capita availability
222

of rice from regional production increased by about 11 percent over the last two
decades. The growth in the productivity of land in rice cultivation (yield rate) was
2.3 percent per year, similar & the performance of Asia as a whole during this period
(Hossain and Laborte, 1994). But the absolute level of yield was only 2.2 tons per
ha, compared to 5.8 tons in China, 4.5 tons in Indonesia and 2.8 tons in Bangladesh
where the production conditions are similar. Within India, the rice yield has reached
4.8 tons in Punjab (in the North) and 4.6 ton in Tamil Nadu (in the south), and in
several districts in these two States farmers get more than 5.5 tons per hectare (GOI,
1993).
The production performance was also uneven across the states in Eastern
India (Table 1). The growth in production was impressive in Uttar Pradesh,
satisfactory in West Bengal, but dismal in Bihar, Orissa and Assam. The annual
growth in rice production was about 5.3 percent in eastern UP, and 2.7 percent in
West Bengal. Nearly four-fifths of the growth was due to the increase in rice yields
made possible through the diffusion of the modern high yielding varieties. In
Madhya Pradesh rice production barely managed to keep pace with the growth of
population; respectively 2.3 and 2.4 percent per year. In Bihar and Orissa, the growth
in rice production was only about 1.6 percent per year. In Orissa, there was an
absolute decline in the area harvested with rice, while rice yield increased by 1.9
percent per year.
Since the land frontier has been closing, the future growth in agricultural
production in the region must come from the growth in productivity of land. The
growth in rice yield was higher than population growth only in Uttar Pradesh and
West Bengal. In other states the productivity growth was less than two percent.
Assam with a growth in yield at less than one percent per year had the most dismal
performance in raising productivity.
An important characteristic of the rice sector in Eastern India is a large year
to year fluctuation in rice production due to unreliable monsoons. For some years
the production shortfall can be astronomical leading to severe food insecurity for
the small and marginal farmers. For example, in Bihar the production shortfall was
67 percent in 1966, 35 percent in 1979, and 24 percent in 1987 compared to the
preceding year. In Madhya Pradesh, the harvest failure was 50 percent in 1965, 58
percent in 1979 and 38 percent in 1986 and 1987. Farmers may suffer from both
floods and droughts in the same season because of heavy early season rains and
prolonged droughts during the end of the monsoon. This makes rice cultivation a
risky venture for majority of the farmers who operate at a subsistence level.
Eastern India experienced considerable rural-urban migration of population
(Table 1). The average population growth rate over the last two decades was 2.2 percent
per year, while the rural population grew at 1.8 percent. The information presented in
Table 1 suggests that the rural-urban migration is partly related to the performance of
the rice sector. The growth in rural population was almost the same as the total
population in UP and West Bengal which experienced an impressive growth in rice
production. In Bihar and Orissa which had sluggish growth, 0.3 to 0.4 percent of the
people migrated every year to urban areas over the last two decades. The rate of
223

migration from Madhya Pradesh was also substantial. An improved performance

of the rice economy can contribute to a reduction in rural-urban migration and help
ease pressures on the overburdened cities and urban areas.
Table 1.

Growth in rice production and productivity in Eastern India, by states,

1970-91.
(Percent per year)
Rice growth

States

Area

Yield

Population
growth

Production

Rural

Total

Assam

0.94

0.92

1.86

n.a.

2.1

West Bengal

0.45

2.27

2.72

2.1

2.2

Bihar

0.04

1.55

1.59

1.8

2.1

-0.34

1.88

1.54

1.4

1.8

0.62

1.68

2.30

1.5

2.4

Uttar Pradesh

1.03

4.28

5.31

1.9

2.1

Eastern India

0.44

2.27

2.71

1.84

2.2

Orissa
Madhya

Pradesh

Source :

IRRI Rice Statistics Database.

Note

The growth rates are obtained by estimating semi- logarithmic trend lines on the time
series data.

In order to assess the regional variation in growth, we also estimated the

production performance at the district level. The districts of Assam were not included
in this anslysis due to non- availability of data for many of the variables considered
in this study. Many districts have been subdivided into smaller districts since the
mid 1970s. Since the data at the subdivided district level was not available for the
earlier period, the estimates are obtained at the greater district level. The growth
rates are estimated for the 1970-87 period, as the data for the district level was not
available for more recent years at the time of the analysis.
The estimated growth in rice and cereal production at the district level are
plotted in Figure 1. The extremely variable performance across districts is clearly
shown by the figure. Nearly 10 percent of the districts had experienced negative
growth, while 20 percent of the districts achieved more than six per cent rate of
growth, which must be regarded as highly impressive by any standard. Almost half
of the districts achieved a rate of growth of less than three percent year which is
needed to meet the population and income growth induced demand for foodgrains
in the region . An important point to note from the Fig. 1 is the highly positive
correlation between the growth in rice and total cereal production. This signifies the
dominant role of rice in achieving food security in Eastern India.
224

PRODUCTION PERFORMANCE FOR AGROECOREGIONS

To what extent can the diverse performance across districts be attributed to
the agroclimatic and environmental factors?. To address this question, we have
classified the districts into relatively homogeneous agroecological zones (AEZ) and
studied the variations in production performance across and within each zone. If
agroecological factors are dominant in influencing rice production we would expect
the growth rate to vary across the AEZs, but little variation across districts within
each AEZ.
Indian scientists have classified India into 21 relatively homogeneous
agroecological regions (Sehgal et al., 1990). The major factors considered in
demarcating the zones are water balance and moisture availability, which determine
the length of growing period (LGP), elevation of the land, soil characteristics, and
the temperature regime as determined by the geographic location of the district.
Most of the districts in Eastern India fall under the following five agroecological
regions :

Identification
of the region
9

Agroecological region

Northern Subhhumid
with alluvial soils

Land area
(million ha)
Plains

12.2

12

Eastern Subhumid Plateau with red

and yellow soils (Chhatisgarh)

13.2

13

Eastern Subhumid Plateau with red

loamy soils (Chhotanagpur)

27.8

14

Eastern Subhumid Plains with alluvial soils

16

Eastern Humid Plains with alluvial soils

9.3
11.7

The location and characteristics of the zones can be seen from Figs. 2 to Fig.
6, which are reproduced from Sehgal et al. (1990).
Table 2 reports the estimates of growth rates in cereal production, and the
sources of growth in the production of rice for the five agroecological regions. The
production performance is impressive for the three AEZs in the subhumid and
humid plains (the Gangetic plain), which increased cereal production at or above
3.5 percent per year over the 1970-87 period. The production of non-rice crops
increased at a faster rate than rice, but the growth in rice production itself surpassed
population growth.

225

Fig. 1.

Growth in cereals production Vs growth in rice production, 1970-87.

The impact of technological progress in rice cultivation can be assessed from

the growth in rice yield. For Northern subhumid plains (AEZ 9) and Eastern humid
plains (AEZ 16), rice yield increased at 2.8 percent per year, a record superior to that
for Asia as a whole (Hossain and Laborte, 1994). In the Eastern subhumid plains
(AEZ 16, North Bihar), nearly one-fourth of the growth on rice production was due
to area expansion, but the absolute level of the growth in rice yield was respectable
at 2.1 percent per year. The Chhotanagyur and Chhatisgarh plateau performed
poorly. It will be noted from the standard errors of the estimates (Table 2) that the
growth in rice yield was fairly stable in the Eastern subhumid and humid plains,
but was highly variable in the Chhatisgarh and Chhotanagpur plateau (AEZ 12) and
in the Northern subhumid plains. The variation in the level and the growth in rice
yield across districts in different agroecological zones can be reviewed from Tables
3 and 4. Yield levels are fairly low. Only in two out of 60 districts in the region, the
average yield for the 1985-87 period was higher than 3.0 tons per hectare, and in 17
districts the yield was less than 1.5 tons per hectare. The majority of the districts in
226

the Eastern humid plains, and Northern subhumid plains had yields over 2.0 tons
per hectare, while most of the districts in the other three AEZs had yields below
that level. The important point to note is that there is a large variation in the level
of yield across districts in all five agroecological zones.
Table 2.

Growth of foodgrains and rice production in Eastern India,

agroecological zones, 1970-87. (Figure in percent per year)

Major

Rice

by

Cereals
production

agroecoregions
area

Yield

Production

Northern Subhumid Plains

with alluvial soils

0.72
(0.29)

2.68
(0.97)

3.43
(1.08)

4.32
(0.62)

12

Eastem Subhumid Pleateau

with red and yellow soils
(Chhatisgarh)

0.54
(0.05

1.68
(1.11)

2.22
(1.13)

2.24
(1.06)

13

Eastern Subhumid Pleateau

with red loamy soils
(Chhotanagpur)

-0.28
(0.20)

1.21
(0.65)

0.92
(0.81)

1.03
(0.75)

14

Eastern Subhumid Plain

with alluvial soils

0.72
(0.19)

2.05
(0.47)

2.78
(0.53)

3.47
(0.36)

16

Eastern Humid plains with

alluvial soils

0.45
(0.19)

2.76
(0.40)

3.22
(0.51)

3.45
(0.48)

Note:

Estimated by fitting semi-logarithmic trend lines with times series data

available at the district level.
Figures within parentheses are standard errors of the estimates. But the
record is dismal for the Chhotanagpur plateau where cereal production
grew at only 1.0 percent per year and rice production at 0.9 percent.
The Chhatisgarh plateau managed to keep pace with the growth of
population with a 2.2 percent per year growth in both rice and cereal
production. But since the region is characterised by widespread poverty,
the income growth induced demand for rice must be significant, and
hence the supply of rice has not been able to keep pace with demand.

The performance with respect to the growth of rice yield also varied widely
across districts. In 19 of the sixty districts under study, the growth in rice yield was
more than 3.0 percent per year; while in nine districts the growth rate was negative,
and most of them are located in the Chhotanagpur plateau. Again, the inter-district
variation in the growth of rice yield was large for each of the AEZ's.
227

Fig. 2.

228

Northern sub-humid plain with alluvial soils.

Fig. 3.

Eastern sub-humid plateau with red and yellow soils.

229

Fig. 4.

230

Eastern sub-humid plateau with red loamy soils.

Fig. 5.

Eastern sub-humid plains with alluvial soils.

231

Fig. 6.

232

Eastern humid plains with alluvial soils.

We conclude from the findings that agroecological factors alone cannot explain
the diverse production performance of the rice sector across districts in Eastern India.

IMPACT OF AGROECOLOGICAL AND SOCIOECONOMIC FACTORS

We used a multivariate regression analysis to assess the importance of
different agroecological and socioeconomic factors in explaining the variation in
the level and the growth of rice yield across districts in Eastern India.
Table 3.

The variation in rice yield by agoecological zones, Eastern India 1985-86.

Average rice yield (t / ha)

Agroecological

zones

< 1.5

Northern Subhumid Plains with alluvial soil

Eastern Subhumid Plateau
with red and yellow soils
Eastern Subhumid Plateau
with red loamy soils
Eastern Subhumid Plains with alluvial soil
Eastern Humid Plains with alluvial soil
Total
(Percent)
Source :

1.52.0

2.03.0

3.0 &
over

Total
10
8

10

3
3

3
5

12
10

17
28.3

23
38.3

18

60

30.3

3.3

100.0

20

Estimated from IRRI Rice Statistics Database.

The socioeconomic factors considered in the analysis are the average size of
the farm in the district, the pattern of distribution of land (measured by the share
of landless agricultural laborers and tenants in total farm population), the incidence
of poverty (measured by the per capita calorie intake of the population for the base
period), and the economic capacity of the farmers (measured by the amount of
chemical fertilizer used per hectare of land in the district). The smaller the size of
farm, the higher the pressure of population on land and hence higher would be the
need to increase foodgrain production. But the size of farm also determines the
availability of labour input and its opportunity cost to the farm household. Small
farms may use more labor and hence achieve higher levels of productivity than large
farms (Berry and Cline, 1979). Small farms may however be economically unable to
afford relatively capital intensive modern technology, and hence the growth in rice
yields may be slower in districts with smaller farm size. The effect of farm size on
productivity would depend on the relative strength of these two oppositive forces.
The unequal distribution of land ownership and the high incidence of tenancy may
provide disincentives to production, and hence is expected to have a negative impact
233

on the growth of rice yield. The subsistence pressure on the growth of rice production
is expected to be captured by the incidence of poverty at the base period.
Table 4.

The variation in growth in rice yield by agroecological zones, Eastern

India, 1970-87.

Agroecological
zones

Annual growth in rice yield (%)

Negative

Northern Subhumid Plains

with alluvial soil
Eastern Subhumid Plateau
with red and yellow soils
Eastern Subhumid Plateau
with red loamy soils
Eastern Subhumid Plains
with alluvial soil
Eastern Humid Plains
with alluvial soils
Total
(Percent)

Up to
1.0

1-0
2.0

2.03.0

Over
3.0

Total

10

20

12

10

11

12

15

13

60

15.0

18.3

20.0

25.0

21.7

100.0

Source: Estimated from IRRI Rice Statistics Database.

The nonavailability of data at the district level was a major constraint to

incorporate in the analysis important agroecological factors that affect rice
production. Climate is an important factor estimated by the average annual rainfall
in the district. The proportion of upland rice area was included to capture the effect
of land elevation. The data on this variable have been taken from Huke and Huke
(1982). For lowlands, farmed control over water supply would be an important
variable affecting the risk of adoption of new technology and investment in modern
agricultural inputs. The percent of rice area under irrigation was included to capture
this effect. We considered flooding depth as another variable affecting rice
production, but it was found highly correlated with the incidence of irrigation and
hence was dropped in the final estimation of the equation.
There are many other bio-physical factors that would affect agricultural
production. Particularly important are soil types, temperature, and pressure of pests
and diseases. Data on these variables are not available at the district-level. In order
to capture the effect of some of the variables, we used dummy variables using the
location of the districts in different agroecological zones. The Eastern humid plains
zones was taken as control (with value o) for assessing production conditions in
other agroecoregions (with value 1).
234

The regression coefficients were estimated using the ordinary Leasts square
method. The results are reported in Table 5.
Table 5.

Factors affecting variation in the level and growth of rice yields across
districts: Regression estimates.
Rice yield

Growth of rice yield

(% / year)

(kg/ha)
Explanatory variables
1
Farm size (ha)
Incidence of landlessness and tenancy
(ratio of total workers)
Per capita energy intake at base
period (10 2 calorie)
Fertilizer use
(kg of materials/ha)
Area under irrigation
(ratio of total area)
Area under upland rice
(ratio of total area)

Eq(1)

Eq(2)

Eq(3)

Eq(4)

-100
(1.15)

-103
(-1.10)

-0.61
(-1.66)

-0.62
(-1.62)

-73
(-0.20)
121
(1.32)

-96
(-0.25)

-3.67**
(-2.37)

-3.11**
(1.95)

77
(0.79)

-0.37
(-0.95)

-0.51
(1.29)

3.3**
(2.97)
855**
(3.26)
314
(0.65)

0.011**
(2.32)
1183**
(4.62)

0.95
(0.86)

2.04*
(1.94)

272
(0.52)

6.78**
(3.30)

6.64**
(3.10)

Annual rainfall (cm)

Agroecoregion dummy
(Eastern Humid Plain =0)

-3.62**
(-2.76)

46.6**
(-3.41)

0.15**
(-2.72)

-0.19**
(-3.32)

Northern SubHumid Plains

-814**
(-3.92)

-1054**
(-5.12)

-0.41
(-0.57)

-1.290
(-1.42)

Eastern SubHumid Plains

-909**
(-5.20)

-1420**
(-6.79)

-1.30
(-1.76)

-2.07**
(-3.08)

Chhotanagpur Plateau

-510**
(-2.46)

-867**
(-4.79)

-0.82
(-0.94)

-2.00**
(2.69)

Chhatisgarh Plateau

-620**
(2.63)

-976**
(-4.45)

-0.54
(-0.54)

-1.72*
(-1.92)

Constant term

2.26**
(5.77)

2.82**
(7.60)

5.29*
(3.19)

7.12
(4.69)

0.77

0.73

0.61

0.56

R2
Note

Figures in parentheses are estimated 't' values. ** denotes that the regression coefficient is
statistically significant at 5% probability level and l at 10 percent.

235

The rice yield is found to be negatively correlated with the amount of rainfall.
The districts with higher rainfall are also found to have slower growth in rice yield.
The regression coefficient for this variable is statistically highly significant. The
results suggest that poor drainage and flooding is a major constraint to increase in
rice yield in Eastern India.
As expected, irrigation is an important factor affecting rice production. The
conversion of rainfed to irrigated land is found to increase rice yield by about 1.2
tons per hectare. Irrigation contributes to the growth in rice yield partly through
facilitating adoption of modern high yielding varieties and larger use of chemical
fertilizers. When the fertilizer variable is included in the model along with the
irrigation variable, the value of the regression coefficient for irrigation is substantially
reduced (Table 5). In fact, in the equation on growth of rice yield, the irrigation
coefficient becomes statistically insignificant as the fertilizer variable picks up the
effect of irrigation. The value of the regression coefficient of fertilizer suggests that
at the farmers level, one kg of additional use of fertilizer (materials) increases rice
production on the margin by 3.3 kg.
Other factors unchanged, the average rice yield is found to be higher in
districts with larger proportion of area under uplands. The positive and statistically
significant regression coefficient for this variable in equations 3 and 4 (in Table 5)
suggests that these districts achieved higher rates of growth in rice yield than the
districts with larger proportion of rainfed lowlands. This finding is contrary to the
prevailing notion that the progress in increasing rice productivity has been extremely
limited in the uplands. This issue needs further investigation.
The association between the level of rice yield and the per capita energy intake
is positive but statistically insignificant. The positive relationship suggests that the
causality may be working the other way - the higher the rice yield, the higher the
per capita energy intake, ie., the lower the incidence of poverty. This result signifies
the importance of increasing rice yield for achieving food secutity. However, the
districts with higher levels of food security have achieved slower growth in rice
yields, as shown by the negative coefficient of the energy intake variable in the
equations for the growth in rice yields.
The districts with smaller size of farms have achieved higher levels of rice
yields and also have higher rates of growth in yields. This is contrary to the notion
that small farms cannot particitate in the adoption of input-intensive technology as
much as the large farms (Griffin, 1974; Pears, 1980). The finding rather supports the
inverse farm size-productivity relationship documented by large number of
micro-studies for South Asia.
The regression coefficient of the landlessness and tenancy variables is negative
and statistically significant in the equation for the growth in rice yields. The value
of the coefficient suggests that the districts with more skewed distribution of land
have lower rice yields and had slower growth in yields than the districts with
relatively less unequal land distribution. This finding points to the importance of
land reforms for providing incentives to growth in agricultural production.
236

The dummy variables for the agroecoregions are statistically highly significant
in the equations for the level of rice yield. The results suggest that there are many
other bio-physical variables which could not be included in the analysis have had
impact on rice yield. Compared to the Humid Plains (mostly in West Bengal) rice
yield are substantially lower not only in the Chhotanagpur and Chhatisgarh Plateau,
but also in the Subhumid Plains in the Upper Ganges Delta. The growth in rice yield
was significantly lower in the two plateau regions and also in Eastern Subhumid
Plains compared to the Humid Plains, even after accounting for differences in
rainfall, irrigation, land levels and the land tenure situation, (eq (4) in Table 5). But
when fertilizer variable is included, the value of the coefficients of the regional
dummy variables declines and becomes statistically insignificant (eq 3). Presumably,
the negative effect of the biophysical constraints in these agroecoregions is partly
offset by larger use of fertilizer input.
What are the factors that influence the use of chemical fertilizers? To address
this issue we conducted another multi- variate analysis relating fertilizer use across
districts to some of the agroecological and socioeconomic variables mentioned above.
We added the standard error of the estimated growth in yield as another explanatory
variable to represent a measure of risk involved in rice cultivation which is a common
problem in Eastern India. The results can be reviewed from Table 6.
Fertilizer use is found to be highly negatively correlated with the instability
in rice production and the relationship is statistically significant. Districts with
more unstable growth in rice yield have lower levels of use of chemicals fertilizers.
As expected, the relationship between irrigation and fertilizer use is positive and
statistically significant. Better water control reduces the risk in rice cultivation
and encourages farmers to apply more fertilizer and adopt fertilizer responsive
varieties.
The coefficient of the farm size variable is negative. It indicates that districts
with smaller farm size have higher intensity of fertilizer use, contrary to a priori
expectation. When the regional dummy variables are introduced the negative
coefficients of farm size and the instability in rice yield turn out to be statistically
insignificant, presumably due to high levels of collinearity between these two sets
of variables. Thus, the lower levels of fertilizer use in the subhumid plains and
plateau regions compared to the humid plains is partly due to larger size of farms
with low subsistence pressure and the higher risks in rice cultivation (Table 2).
In a recent study on research prioritization for eastern India, major technical
constraints to increase rice production are reported as drought, weeds, yellow stem
borer, acid soils, bacterial leaf blight, lodging, blast, submergence, zinc deficiency
and, army worm in order of importance (Widawsky and O Toole, 1990). The study
estimates that these top 10 problem areas account for 60 per cent of the losses due
to technical constraints that contribute to nearly 45 per cent of the gap between the
average farm level yield and the yield obtained from on-farm experiments under
optimal environmental conditions. The study on agroecological zoning for India
mentions flooding, poor drainage and water-logging as key constraints to
agricultural production in the humid and subhumid plains in the Gangetic Delta;
237

and drought, problem soils and soil erosion as key constraints in the Chhotanagpur
and Chhatisgarh plateau (Sehgal et al., 1990).
Table 6.

Factors affecting variation in fertilizer use across districts in Eastern

India: Regression Estimates.

Factors

Fertilizer use (kg/ha)

Eq(1)

Eq(2)

Farm size (ha)

-36.15**
(-3.79)

-8.04
(-0.74)

Instability in yield (SE)

-36.17**
(-2.47)

-13.15
(-0.90)

-1.00
(0.58)

-3.77*
(2.34)

97.13**
(3.33)

83.95**
(3.08)

33.78
(0.50)

-46.12
(-0.77)

Annual rainfall (cm)

Area irrigated (ratio of
total rice area)
Area under upland rice
(ratio of total rice area)
Agroecoregiondummy (Humid Plains = 0)
Northern
Eastern

Subhumid
Subhumid

Chhotanagpur
Chhatisgarh
Constant
R2
Note

-72.0**
(2.93)

Plains

-68.9**
(3.26)

Plains

-112.2**
(5.08)

Plateau

-103.8**
(-3.52)

Plateau
term

140.36**
(3.30)

207.9**
(5.22)

0.56

0.71

Figures within parentheses are estimated 't' value of the regression coefficents. ** denotes
statistical significance at 5 percent level.

Mechanism of tolerance to drought particularly at the anthesis stage, is crucial

to increased productivity not only in the plateau regions but also in the humid and
238

subhumid plains which suffer frequent late season moisture stress. Submergence
tolerance is crucial not only for areas with problems of flooding and poor drainage
(in Ganges Delta) but also for the Chhatisgarh and Chhotanagpur plateau where
rainfall suddenly picks up with frequent heavy rains for a number of consecutive
days during the months of July and August and then the monsoon abruptly ends
in the month of September (See Figs. 3 and 4). Elongating varieties tend to lodge
after the flood water recede. So, semi-dwarf varieties with abilities to short- term
submergence could prove beneficial to large areas of Eastern India.
Development of resistant cultivars to prolonged moisture stress and short
period submergence can stabilize yields and reduce the risk in rice cultivation, and
thereby induce low-income farmers to adopt input-intensive technologies. The
rainfed Lowland Rice Research Consortium is on the right track in emphasizing
stress physiology as the priority area of research for this region.

REFERENCES
Berry A R, Cline W R (1979) Agrarian Structure and Productivity in Developing Countries. Baltimore: Johns
Hopkins University Press.
Dev S, Suryanarayana M H, Parikh S (1992) "Rural poverty in India: Incidence, Issue and policies", Asian
Development Review, 10 (1) :35-66.
Fertilizer Association of India (1991) Fertilizer Statistics, 1990-92. New Delhi, India.
Government of India (GOI) (1993) Agricultural Situation in India. September 1993 (earlier issues), Directorate
of Economics and Statistics, Ministry of Agriculture, New Delhi, India.
Griffin K (1974) The Political Economy of Agrarian Change: An Essay on the Green Revolution. Cambridge:
Harvard University Press.
Hossain M, Laborte A (1994) "Rice Economy in Asia: Recent Progress and Emerging Trends", Food and
Fertilizer Technology Centre Newsletter, 101. Taipei, Taiwan.
Huke R, Huke E (1982) Rice Area by Type of Culture: South, Southeast and East Asia. International Rice
Research Institute, Los Banos, Philippines.
Huke R, Huke E (1990) Human Geography of Rice in South Asia. International Rice Research Institute, Los
Banos, Philippines.
Kumar P, Rosegrant M W (1993) "Productivity and sources of growth in rice in India", paper presented
at the Second Workshop on Rice Supply and Demand Project, International Rice Research Institute and
International Food Policy Research Institute, 13-15 April 1993, Los Banos, Philippines.
Pears A (1980) Seeds of Plenty, Seeds of Want: Social and Economic Implications of the Green Revolution Oxford:
Clarendon Press.
Sehgal J L, Mandal D K , Mandal C, Vadivelu S (1990) Agro-Ecological Regions of India, NBSS Publication
No. 24, National Bureau of soil Survey and Land Use Planning, Nagpur, India.
Widawsky, D, OToole J C (1990) "Prioritizing the Rice Biotechnology Research Agenda for Eastern India",
The Rockefeller Foundation.

239

Printed at Prakash Packagers, 257-Golaganj, Lucknow.